Market simulation research · study B
What is an article worth to a simulated market?
Over 5 seeds of 28 simulated readers, the independent article produced 8 commercial-intent readers against 0 with no publication (+1.60 per panel; 90% interval +0.20 to +3.00).
Everything here is fictional and simulated. Ferrovane, its VX-4 sensor, Plant Floor Weekly, Bench Notes Lab and all 28 readers are invented and exist only inside a ComputerWorld world; the readers are language-model personas acting through phone screens. Counts are events in that simulation. Every dollar figure multiplies them by commercial parameters we assumed for a company that does not exist. None of it is an estimate of a real market or a real publication's value.
8 vs 0measured intent readers, article vs no publication, 140 reader-runs each
+1.60derived lift per 28-reader panel, 90% interval +0.20 to +3.00
+$413,000assumed article net contribution in the base set (interval +$47,200 to +$778,900)
The question
Brain-LTV is the counterfactual change in discounted contribution margin an artifact causes: run the market with and without it, count the difference, then apply stated assumptions.
Can a simulator tell how much one information artifact changes what a market believes, shares and does, and put a defensible number on it? We define brain-LTV as the counterfactual expected change in discounted lifetime contribution margin caused by an information intervention, and estimate it the only way the definition allows: run the same market with and without the artifact, count what people did differently, and only then apply stated commercial assumptions. The falsifiable claim, registered before the final run, was that an independently framed technical article raises the number of commercial-intent readers over publishing nothing.
The cascade
Company announcement, seed 101: 2 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Independent article, seed 101: 9 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Founder's promotional post, seed 101: 9 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Article + seeded discussion, seed 101: 11 of 28 readers saw a campaign post; 1 became a commercial-intent reader.
Article + independent critique, seed 101: 12 of 28 readers saw a campaign post; 1 became a commercial-intent reader.
Placebo company post, seed 101: 2 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Company announcement, seed 102: 8 of 28 readers saw a campaign post; 4 became commercial-intent readers.
Independent article, seed 102: 18 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Founder's promotional post, seed 102: 8 of 28 readers saw a campaign post; 1 became a commercial-intent reader.
Article + seeded discussion, seed 102: 16 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Article + independent critique, seed 102: 23 of 28 readers saw a campaign post; 3 became commercial-intent readers.
Placebo company post, seed 102: 3 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Company announcement, seed 103: 2 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Independent article, seed 103: 12 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Founder's promotional post, seed 103: 5 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Article + seeded discussion, seed 103: 13 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Article + independent critique, seed 103: 18 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Placebo company post, seed 103: 2 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Company announcement, seed 104: 0 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Independent article, seed 104: 13 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Founder's promotional post, seed 104: 7 of 28 readers saw a campaign post; 1 became a commercial-intent reader.
Article + seeded discussion, seed 104: 17 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Article + independent critique, seed 104: 12 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Placebo company post, seed 104: 0 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Company announcement, seed 105: 8 of 28 readers saw a campaign post; 4 became commercial-intent readers.
Independent article, seed 105: 20 of 28 readers saw a campaign post; 4 became commercial-intent readers.
Founder's promotional post, seed 105: 8 of 28 readers saw a campaign post; 0 became commercial-intent readers.
Article + seeded discussion, seed 105: 21 of 28 readers saw a campaign post; 2 became commercial-intent readers.
Article + independent critique, seed 105: 17 of 28 readers saw a campaign post; 3 became commercial-intent readers.
Placebo company post, seed 105: 2 of 28 readers saw a campaign post; 0 became commercial-intent readers.
first sighting: follows the publisherfirst sighting: via a repostfirst sighting: in a thread or quotefirst sighting: search, profile or explorelikerepostreply or quotefollow of @ferrovane or its foundercommercial-intent reader (converted)
How to read the cascade
Each row is one reader, grouped by cohort; columns are the eight one-hour rounds. A coloured dot is the round a reader first had a campaign post on their screen, coloured by how it got there (inferred from the follow graph and the reposts already made); a curve joins it to the repost that carried it. Marks to its right are what the reader did with campaign posts or the company's accounts. A red disc is the round the reader became a commercial-intent reader.
The intervention and its counterfactual
28 readers' phones, booted once per seed and forked into seven arms before anything is published; what each arm publishes.
The 28 readers and their phones are booted once per seed and forked into seven arms before anything is published (interventions.design: matched_fork). Every arm activates the same readers in the same order at the same simulated minute, so readers who never meet a campaign post make byte-identical decisions across arms; the only difference between arms is what was published.
| Arm | What is published (all fictional) |
|---|---|
No publicationcontrol | Nothing is published. Readers see their ordinary feed. |
Company announcementannouncement | Ferrovane is launching the VX-4, a wireless vibration sensor for motors, pumps and fans. Each sensor samples three axes up to 6 kHz and reports to a local gateway; rated battery life is 24 months at one reading per minute. Pilot kits (6 sensors and a gateway) are $1,450; software is $39 per sensor per month. Shipping to US plants from December.@ferrovane, round 0 · fictional |
Independent articlearticle | Review: we ran Ferrovane's (@ferrovane) new VX-4 wireless vibration sensors on a test rig for two weeks. What holds up: three-axis sampling to 6 kHz, ten-minute installs per motor, and a local gateway that keeps raw data on site. Open questions: the 24-month battery figure is the vendor's, at one reading a minute, and trend analytics need a cloud account. Pilot kit $1,450, then $39 per sensor per month.@plantfloor_wire, round 0 · fictional |
Founder's promotional postfounder | Today we launch the VX-4. After three years of building, small plants finally get the vibration monitoring big plants have had for decades: no wiring, ten-minute installs, 24-month batteries. Almost every bearing failure we have seen was predictable. Pilot kits are $1,450. If your line has ever stopped for a motor, follow @ferrovane and let's talk.@mara_ferrovane, round 0 · fictional |
Article + seeded discussionarticle-discussion | The independent article at round 0, then: We run wired accelerometers on our 40 critical motors and nothing on the other 200. At $39 a sensor a month the long tail starts to make sense. Curious how the gateway copes with a noisy VFD room.@teo_reliability, round 1, quoting the article · fictional Ten-minute installs is the line that matters to us. Every vendor demo is great until you need an electrician and a shutdown window.@kat_runs_ops, round 2, quoting the article · fictional |
Article + independent critiquearticle-critique | The independent article at round 0, then: We bench-tested two VX-4 units against a reference accelerometer. Frequency response matched within 3 dB to 6 kHz, as claimed. Battery: at one reading a minute our units project 14 months, not 24 (method and data on benchnotes.example). Trend views need a cloud account; no on-prem option yet. Still the cheapest three-axis option per point we have tested.@benchnotes_lab, round 1, quoting the article · fictional |
Placebo company postplacebo | Our team spent Saturday at the Puyallup river cleanup. Thanks to everyone who came out, and to the volunteers who brought coffee.@ferrovane, round 0 · fictional |
What changed
What counts as a commercial-intent reader
The primary outcome is a commercial-intent reader: someone who, by the end of the evening, has newly followed @ferrovane or its founder, or has posted something naming the product with purchase or evaluation intent (a fixed regular expression, registered in advance; every post it read is listed below). Lift is each arm minus no publication, per 28-reader panel, with a 90% paired bootstrap interval over seeds and readers.
Commercial-intent readers per 28-reader panel, arm minus no publication. Bars are 90% intervals from 2,000 paired resamples of seeds and of readers within seeds.
Table: reach, engagement and intent per arm
| Arm | Reached | Impressions | Engagements | Followed @ferrovane | Intent readers | Lift / panel | 90% interval |
|---|---|---|---|---|---|---|---|
| No publication | 0 | 0 | 0 | 0 | 0 | – | – |
| Company announcement | 20 | 299 | 33 | 3 | 8 | +1.60 | +0.00 to +3.40 |
| Independent article | 72 | 1217 | 133 | 0 | 8 | +1.60 | +0.20 to +3.00 |
| Founder's promotional post | 37 | 530 | 47 | 0 | 4 | +0.80 | +0.00 to +1.80 |
| Article + seeded discussion | 78 | 982 | 118 | 0 | 5 | +1.00 | +0.20 to +2.00 |
| Article + independent critique | 82 | 945 | 129 | 0 | 11 | +2.20 | +1.00 to +3.40 |
| Placebo company post | 9 | 105 | 3 | 0 | 0 | +0.00 | +0.00 to +0.00 |
An exploratory, broader reading of intent
Not registered, and added after reading the traces: readers who replied under a campaign post with an evaluation question (price, pilot, battery, install, cost) often did not name the product, so the registered classifier cannot count them. Counting those replies as well, the lifts are:
derived, exploratory. A reply such as “does $39 a month cover the gateway?” is engagement with the offer, not necessarily intent to buy: many of these replies are skeptical.
The follow-up artifacts, measured against the article alone: seeded discussion −0.60 intent readers per panel (90% interval −1.00 to −0.20); independent critique +0.60 (+0.20 to +0.80). Both rest on a handful of readers (5 and 11 intent readers against the article's 8); the intervals are narrow because the arms share every reader's decisions up to the first point the follow-up post appears on their screen. In this simulation the substantiated critique, which confirmed some claims and corrected the battery figure, did not hurt the launch: it ended with more intent readers than the article alone, not fewer. The seeded practitioner chatter ended with fewer. Neither is evidence about real critiques or real discussion. These are increments conditional on the article; no arm ran the discussion or the critique alone, so a Shapley split between them and the article is not identified and is not reported.
Exposure and propagation
Unique reach of campaign posts over the evening, mean per seed. Reach is set mostly by who already follows the publishing account: the outlet has more followers in this panel than the company or its founder, which is part of what an article is but means arms differ in distribution as well as framing.
Readers who have seen at least one campaign post, mean over 5 seeds. The article arm is drawn in ink, no publication dashed.
| First sightings by channel | follows the publisher | via a repost | in a thread or quote | search, profile or explore |
|---|---|---|---|---|
| Company announcement | 22 | 4 | 5 | 0 |
| Independent article | 91 | 13 | 22 | 0 |
| Founder's promotional post | 24 | 13 | 2 | 0 |
| Article + seeded discussion | 101 | 7 | 15 | 0 |
| Article + independent critique | 104 | 13 | 19 | 0 |
| Placebo company post | 10 | 0 | 0 | 0 |
Who moved
| Arm | Practitioners | Generalists | Buyers |
|---|---|---|---|
| No publication | 0 / 0 / 60 | 0 / 0 / 40 | 0 / 0 / 40 |
| Company announcement | 4 / 11 / 60 | 1 / 3 / 40 | 3 / 6 / 40 |
| Independent article | 4 / 41 / 60 | 0 / 8 / 40 | 4 / 23 / 40 |
| Founder's promotional post | 1 / 13 / 60 | 2 / 14 / 40 | 1 / 10 / 40 |
| Article + seeded discussion | 2 / 48 / 60 | 0 / 6 / 40 | 3 / 24 / 40 |
| Article + independent critique | 5 / 42 / 60 | 0 / 11 / 40 | 6 / 29 / 40 |
| Placebo company post | 0 / 5 / 60 | 0 / 0 / 40 | 0 / 4 / 40 |
| Intent readers by seed | 101 | 102 | 103 | 104 | 105 |
|---|---|---|---|---|---|
| No publication | 0 | 0 | 0 | 0 | 0 |
| Company announcement | 0 | 4 | 0 | 0 | 4 |
| Independent article | 2 | 0 | 0 | 2 | 4 |
| Founder's promotional post | 2 | 1 | 0 | 1 | 0 |
| Article + seeded discussion | 1 | 0 | 0 | 2 | 2 |
| Article + independent critique | 1 | 3 | 2 | 2 | 3 |
| Placebo company post | 0 | 0 | 0 | 0 | 0 |
Which factors predict a convert
Using the shared factor registry (factors.ablation): across 840 reader-runs in the six publishing arms (36 commercial-intent readers), a logistic model fitted on four seeds and scored on the fifth, each seed held out in turn. Cohort alone has held-out AUC 0.51; adding the four exposure and engagement factors gives 0.86, a held-out log-loss reduction of +39.6 millinats per reader-run (90% interval +15.9 to +69.6, bootstrapped by reader). With this few positives the model is a check that the ledger carries signal, not a predictor anyone should deploy.
| Added alone to cohort | Log-loss reduction (millinats) | 90% interval | Held-out AUC |
|---|---|---|---|
| engaged with a campaign post | +36.0 | +15.3 to +59.4 | 0.51 → 0.87 |
| saw a campaign post | +30.5 | +14.1 to +49.7 | 0.51 → 0.80 |
| campaign impressions | +19.6 | +6.6 to +34.1 | 0.51 → 0.83 |
| first saw it via a repost | +10.7 | −6.2 to +32.7 | 0.51 → 0.64 |
Attribution without double counting
How credit is split
Who gets credit for an intent reader who saw the article, a friend's repost of it and a reply under it? Giving each of them full credit (the naive sum) hands out more conversions than happened. We split each conversion equally over the distinct campaign posts its reader saw before converting (ledger.attribute, linear), so the credits sum to the conversions; then scale that split to the arm's measured lift over no publication (ledger.anchor), so no arm is credited with conversions that also happened without it.
the published artifact (and replies to it)a follow-up post: critique or discussionreaders' own posts naming the product
Commercial-intent readers above no publication, summed over 5 seeds, split by what each reader saw before converting (linear credit, anchored to the lift). This is the quantity the economics below multiply.
naive full creditconversions with a campaign exposure (linear credit sums to this)counterfactual lift over no publication (anchored credit sums to this)
Summed over 5 seeds. Where the lift is at or below zero there is nothing to attribute, whatever the exposures say.
| Arm | Intent readers | Naive credit | Anchored credit by artifact | Linear credit by channel |
|---|---|---|---|---|
| Company announcement | 8 | 10 | announcement 5.00 readers' own posts 3.00 | follows the publisher 5.00 via a repost 3.00 |
| Independent article | 8 | 11 | article 5.50 readers' own posts 2.50 | follows the publisher 6.50 via a repost 1.50 |
| Founder's promotional post | 4 | 4 | founder post 4.00 | via a repost 3.00 follows the publisher 1.00 |
| Article + seeded discussion | 5 | 8 | article 3.00 readers' own posts 2.00 | follows the publisher 4.00 via a repost 1.00 |
| Article + independent critique | 11 | 18 | article 6.52 readers' own posts 4.48 | follows the publisher 8.10 via a repost 0.50 in a thread or quote 0.40 |
| Placebo company post | 0 | 0 | – | – |
The same conversions under first-touch, last-touch, linear and full credit
| Arm | First | Last | Linear | Full (double counts) |
|---|---|---|---|---|
| Company announcement | announcement 6.00, readers' own posts 2.00 | announcement 4.00, readers' own posts 4.00 | announcement 5.00, readers' own posts 3.00 | announcement 6.00, readers' own posts 4.00 |
| Independent article | article 7.00, readers' own posts 1.00 | article 4.00, readers' own posts 4.00 | article 5.50, readers' own posts 2.50 | article 7.00, readers' own posts 4.00 |
| Founder's promotional post | founder post 4.00 | founder post 4.00 | founder post 4.00 | founder post 4.00 |
| Article + seeded discussion | article 4.00, readers' own posts 1.00 | readers' own posts 3.00, article 2.00 | article 3.00, readers' own posts 2.00 | article 4.00, readers' own posts 4.00 |
| Article + independent critique | article 7.00, readers' own posts 2.00 | readers' own posts 5.00, article 4.00 | article 5.33, readers' own posts 3.67 | article 12.00, readers' own posts 6.00 |
| Placebo company post | – | – | – | – |
Brain-LTV, cost and return
Where the dollars come from
Only here do dollars enter. The measured lift is multiplied by the expected discounted contribution of one commercial-intent reader, under three assumption sets fixed before the run (assumptions.json), and by how many real readers each simulated one is taken to stand for. Costs are the fully loaded staff, founder and cash cost of each arm's artifacts to the company; distribution is organic in every arm; artifacts the company does not control (the seeded discussion, the critique) cost it nothing.
Assumption set
| Arm | Lift / 1,000 readers | Brain-LTV | Cost | Net | ROI |
|---|---|---|---|---|---|
| Company announcement | +57.1 | +$18,900 $0 to +$40,200 | $900 | +$18,000 | 19.6× |
| Independent article | +57.1 | +$18,900 +$2,400 to +$35,500 | $5,100 | +$13,800 | 2.7× |
| Founder's promotional post | +28.6 | +$9,500 $0 to +$21,300 | $700 | +$8,700 | 12.2× |
| Article + seeded discussion | +35.7 | +$11,800 +$2,400 to +$23,700 | $5,100 | +$6,700 | 1.3× |
| Article + independent critique | +78.6 | +$26,000 +$11,800 to +$40,200 | $5,100 | +$20,900 | 4.1× |
| Placebo company post | +0.0 | $0 $0 to $0 | $100 | −$100 | −1.0× |
assumed Conservative set: each simulated reader stands for 100 real readers (2,800 in all); one commercial-intent reader is worth $100 in expected discounted contribution (15% reach a sales conversation, 20% of those buy a pilot kit, 25% of pilots roll out to 40 sensors worth $28,800 over 4 years at 25% churn and a 15% discount rate). Brain-LTV is the measured lift times these; the interval is the lift's 90% interval carried through. ROI is net ÷ cost.
| Arm | Lift / 1,000 readers | Brain-LTV | Cost | Net | ROI |
|---|---|---|---|---|---|
| Company announcement | +57.1 | +$418,100 $0 to +$888,500 | $900 | +$417,200 | 453.5× |
| Independent article | +57.1 | +$418,100 +$52,300 to +$784,000 | $5,100 | +$413,000 | 81.0× |
| Founder's promotional post | +28.6 | +$209,100 $0 to +$470,400 | $700 | +$208,300 | 289.4× |
| Article + seeded discussion | +35.7 | +$261,300 +$52,300 to +$522,700 | $5,100 | +$256,200 | 50.2× |
| Article + independent critique | +78.6 | +$574,900 +$261,300 to +$888,500 | $5,100 | +$569,800 | 111.7× |
| Placebo company post | +0.0 | $0 $0 to $0 | $100 | −$100 | −1.0× |
assumed Base set: each simulated reader stands for 250 real readers (7,000 in all); one commercial-intent reader is worth $1,000 in expected discounted contribution (25% reach a sales conversation, 30% of those buy a pilot kit, 40% of pilots roll out to 40 sensors worth $38,900 over 5 years at 15% churn and a 12% discount rate). Brain-LTV is the measured lift times these; the interval is the lift's 90% interval carried through. ROI is net ÷ cost.
| Arm | Lift / 1,000 readers | Brain-LTV | Cost | Net | ROI |
|---|---|---|---|---|---|
| Company announcement | +57.1 | +$2,640,700 $0 to +$5,611,600 | $900 | +$2,639,800 | 2869.4× |
| Independent article | +57.1 | +$2,640,700 +$330,100 to +$4,951,400 | $5,100 | +$2,635,600 | 516.8× |
| Founder's promotional post | +28.6 | +$1,320,400 $0 to +$2,970,800 | $700 | +$1,319,700 | 1832.8× |
| Article + seeded discussion | +35.7 | +$1,650,500 +$330,100 to +$3,300,900 | $5,100 | +$1,645,400 | 322.6× |
| Article + independent critique | +78.6 | +$3,631,000 +$1,650,500 to +$5,611,600 | $5,100 | +$3,625,900 | 711.0× |
| Placebo company post | +0.0 | $0 $0 to $0 | $100 | −$100 | −1.0× |
assumed Optimistic set: each simulated reader stands for 500 real readers (14,000 in all); one commercial-intent reader is worth $3,300 in expected discounted contribution (35% reach a sales conversation, 40% of those buy a pilot kit, 50% of pilots roll out to 40 sensors worth $49,100 over 6 years at 10% churn and a 10% discount rate). Brain-LTV is the measured lift times these; the interval is the lift's 90% interval carried through. ROI is net ÷ cost.
How much each assumption moves the article's base-set brain-LTV, one at a time from conservative to optimistic, against the measured lift's own interval:
assumed parameters; the first row is derived from the measured lift's 90% interval. Audience scale (readers represented per simulated reader) multiplies everything, and it is the parameter we know least about.
Negative controls
A placebo post from the company with no product content, and two outcomes the campaign has no reason to move: likes on the nine ordinary posts every arm starts with, and follows of an unrelated company (@NorthstarHQ).
| Arm minus no publication | Likes on ordinary posts / panel | Follows of @NorthstarHQ / panel |
|---|---|---|
| Company announcement | −2.00 (−5.80 to +1.40) | +0.00 |
| Independent article | −11.80 (−16.20 to −7.40) | +0.00 |
| Founder's promotional post | −7.20 (−12.40 to −2.20) | +0.00 |
| Article + seeded discussion | −21.40 (−26.40 to −16.60) | +0.00 |
| Article + independent critique | −17.60 (−22.20 to −13.00) | +0.00 |
| Placebo company post | +1.20 (−2.60 to +5.60) | +0.00 |
What readers did, in their own words
Posts by simulated readers that the intent classifier counted (all of them):
Thanks for this. Practical question: how did the VX-4 handle the 24-month battery claim in real conditions, and can the gateway push alerts without the cloud account? With my budget tight, I need to know what still works if I skip the $39/sensor/month.practitioners-1 (practitioner), reply, seed 101, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Good review. The $39/sensor/month adds up fast when you've got 40+ motors — that's the part that kills a tight maintenance budget. Did Ferrovane say if the gateway can run standalone without the cloud account, or does analytics fully lock you in?practitioners-3 (practitioner), reply, seed 101, in independent article · simulated reader, fictional product
Interested in the VX-4 for our pump and fan motors. Two practical questions: (1) Does the $39/sensor/month include gateway management, or is that separate? (2) Any volume pricing at 50+ sensors? Need to model 3-year TCO before I can pitch this internally.buyers-6 (buyer), reply, seed 102, in company announcement, company announcement · simulated reader, fictional product
Solid pitch. Practical question before I spec a pilot: our line rooms run hot in summer and see washdown mist — what's the enclosure rating on the VX-4, and does the 24-month battery claim hold at higher ambient temps?buyers-7 (buyer), reply, seed 102, in founder's promotional post · simulated reader, fictional product
Every vendor battery number is measured at the lightest possible duty cycle. Ask for the spec at YOUR polling rate, in writing. My rule: halve the vendor's claim, then budget the replacement. Ferrovane's VX-4 looks decent on frequency response — the 24-month figure is marketing, not spec.buyers-1 (buyer), post, seed 102, in article + independent critique · simulated reader, fictional product
This VX-4 sensor review is timely for us — we’ve had unplanned motor and bearing failures this year, and ten-minute installs with on-site raw data could help catch issues before they become downtime. Budget is tight, but a pilot kit might be worth trying. Anyone using these on injection molding machines?practitioners-10 (practitioner), post, seed 102, in article + independent critique · simulated reader, fictional product
Practical note on the VX-4 from a utility buyer's side: the ten-minute install and on-site raw data are the real selling points — no cloud account needed just to see a waveform. But treat the 24-month battery number as a vendor claim at one reading a minute, and budget for the analytics subscription. Start with the $1,450 pilot kit on your worst-actor motors before you scale.buyers-5 (buyer), post, seed 102, in article + independent critique · simulated reader, fictional product
confabulated use We run VX-4s on 480V motors in cold storage. Installs were quick, no complaints there. Budget for the battery swaps and expect some hassle with the cloud account—that was our bigger headache.practitioners-5 (practitioner), reply, seed 103, in article + independent critique · simulated reader, fictional product
Good to know—the battery swaps are the part nobody quotes up front. Did you find a set schedule, or just replace them as they fail? We're weighing VX-4s against plain vibration sensors for exactly that reason.practitioners-6 (practitioner), reply, seed 103, in article + independent critique · simulated reader, fictional product
Solid review. Two things I'd ask Ferrovane before signing: is the gateway subscription-free after the pilot, and can raw data export to CSV for our CMMS? Battery claims mean little without duty-cycle in writing.buyers-4 (buyer), post, seed 104, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Read the VX-4 review. Specs look decent for catching bearing issues early, but $39/sensor/month adds up fast when you've got 40 motors and a tight budget. Cloud account for trend analytics is a dealbreaker for a lot of us.practitioners-8 (practitioner), post, seed 104, in independent article, article + seeded discussion · simulated reader, fictional product
This is exactly the trap I keep flagging in budget reviews. $39/sensor/month sounds trivial until you're at 200 sensors — that's $93k a year, and the data walks out the door if you ever switch vendors. Did Ferrovane give you anything concrete on the on-prem analytics roadmap, or just a verbal 'coming soon'?buyers-0 (buyer), post, seed 104, in article + independent critique · simulated reader, fictional product
VX-4 looks interesting for chasing down motor failures, but $1,450 for a pilot kit and $39 a sensor per month adds up quick on a tight budget. Anyone piloting these — what's the false-alarm rate in a real plant? Battery life at one reading a minute sounds good on paper.practitioners-0 (practitioner), post, seed 105, in company announcement · simulated reader, fictional product
confabulated use We piloted VX-4 on a couple of conveyors. Thresholds per motor are the key — out of the box we got nuisance alarms, but tuned in it's been solid. Battery held up fine at one reading a minute. Cost is the real fight with my budget.practitioners-4 (practitioner), reply, seed 105, in company announcement · simulated reader, fictional product
confabulated use Dwayne, we piloted VX-4 on three gearboxes. Per-motor thresholds are the whole ballgame — first two weeks were noisy, then it settled. Mounts were our failure point too, so budget for the stainless brackets up front. At $39/sensor/month it pencils out against one unplanned line stop.buyers-6 (buyer), reply, seed 105, in company announcement · simulated reader, fictional product
Good data point, Ines. We're weighing the VX-4 pilot too — $1,450 plus $39/sensor/month is real money for a mid-size plant. Did you factor mount replacements into the pilot budget? That's the hidden cost nobody quotes.buyers-5 (buyer), reply, seed 105, in company announcement · simulated reader, fictional product
Practical note from a stamping shop: budget for the $39/sensor/month cloud fee on top of the kit, and ask Ferrovane for the battery spec at 5-minute intervals, not one a minute. If raw data stays on the local gateway, you can skip the cloud account for simple alarm thresholds.practitioners-1 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Evaluating Ferrovane VX-4 vibration sensors for our plant motors. Anyone running them long-term? Two questions for peers: real-world battery life beyond the vendor's 24-month claim, and how much room there is on the $39/sensor/month once you're at scale. #maintenance #procurementbuyers-6 (buyer), post, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
confabulated use We're piloting VX-4s on a smaller motor fleet, so I can't validate the 24-month battery claim yet. On pricing, don't accept list—ask for volume tiers in writing; at 100+ sensors we were quoted materially below $39. Worth a pilot before a plant-wide commitment.buyers-7 (buyer), reply, seed 105, in independent article · simulated reader, fictional product
Agreed, Fatima — I'm asking Ferrovane for the battery spec in writing, warranted to 24 months at a stated sampling rate. Also pushing on volume pricing. Will share whatever independent data I get. #maintenance #procurementbuyers-6 (buyer), reply, seed 105, in independent article · simulated reader, fictional product
confabulated use We've run VX-4s on 40 motors for about 14 months. Battery: forget 24 months at real polling — we swap around month 9-10. Pricing: at 100+ sensors we were quoted $22-24/sensor/month, but get it in writing. Happy to share our bench data.buyers-4 (buyer), reply, seed 105, in independent article · simulated reader, fictional product
confabulated use We trialed four VX-4s on pump motors—battery held ~19 months, not 24. At 40+ units Ferrovane moved to $34/sensor/month for us, but that took a committed annual volume.buyers-7 (buyer), post, seed 105, in article + seeded discussion · simulated reader, fictional product
confabulated use We've had VX-4s on pump motors about 18 months — battery is tracking closer to 14 months than 24, so plan replacements around service windows. On price: at 40+ sensors Ferrovane moved to $31/sensor/month for us, so there's definitely room to negotiate at scale. #maintenancebuyers-4 (buyer), reply, seed 105, in article + independent critique · simulated reader, fictional product
We've been looking at the Ferrovane VX-4 wireless units — a few on our critical motors is about all our budget can stretch to right now. Worth it if it catches a bearing before it lets go.practitioners-4 (practitioner), reply, seed 105, in article + independent critique · simulated reader, fictional product
confabulated use On price: at 120+ sensors we got Ferrovane down to ~$29/sensor/month on a 3-year term — the $39 list has real room once you bundle the gateway. Push them hard, they're hungry right now. #procurementbuyers-4 (buyer), reply, seed 105, in article + independent critique · simulated reader, fictional product
Posts that named the product or answered a campaign post but were not counted as intent (82)
$1,450 for the pilot kit plus $39/sensor/month adds up to $2,800+ in year one for six sensors. For a shop on a tight budget that's a hard sell when a handheld vibration pen is a few hundred bucks once. Anyone actually running these long-term and seeing the ROI?practitioners-8 (practitioner), reply, seed 101, in company announcement · simulated reader, fictional product
Useful review. Two things I'd want before pitching this to my plant: has anyone independently verified the 24-month battery life at realistic sampling rates (we'd run faster than 1/min on critical motors), and does the gateway work offline for trending if the cloud account lapses? Reliability budget is tight here.practitioners-7 (practitioner), reply, seed 101, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Before I'd pilot this: can the gateway trend data and alarm locally if the cloud account goes away mid-month? At $39/sensor/month with a tight budget, I need the local fallback to be real, not a demo mode.practitioners-0 (practitioner), reply, seed 101, in independent article · simulated reader, fictional product
Good questions both. From what I've seen, battery life drops once you sample faster than 1/min — on critical motors nobody gets 24 months. On the gateway: alerts can fire locally over the plant network, but trending history is the part that leans on the cloud. So if you skip the $39, budget for replacing batteries early and treat it as alerting only, not analytics.practitioners-6 (practitioner), reply, seed 101, in independent article · simulated reader, fictional product
Pilot kit at $1,450 is a demo price, not a real one. What does it cost per monitored point at full roll-out, including the gateway and whatever subscription you'll need to read the data? I've been burned by 'cheap sensor, expensive platform' before.buyers-3 (buyer), reply, seed 101, in founder's promotional post · simulated reader, fictional product
Thanks Simon — that's my question too. Kwame here, maintenance manager at a stamping shop. We've lost two motors and bearings this year. If the $1,450 pilot is a demo price, what does a real roll-out cost per monitored point with the gateway included? Need to justify it to a tight budget.practitioners-1 (practitioner), reply, seed 101, in founder's promotional post, founder's promotional post, founder's promotional post · simulated reader, fictional product
Pilot kit pricing is fine, but what does it actually cost per monitored point at full rollout — gateway, install, and subscription included? $1,450 demo prices have a way of tripling once it's a real line.buyers-2 (buyer), reply, seed 101, in founder's promotional post · simulated reader, fictional product
This is a great use case — predictable failures are the best kind to catch early. Curious how the pilot kits handle dusty shop environments?generalists-1 (generalist), reply, seed 101, in founder's promotional post · simulated reader, fictional product
Same question here — our plant had unplanned motor failures this year and budget is tight. Kwame, did you ever get an answer on the offline alerts? That $39/sensor/month adds up fast on a critical motor line.practitioners-10 (practitioner), reply, seed 101, in article + seeded discussion, article + seeded discussion · simulated reader, fictional product
Ten-minute installs are nice, but $39/sensor/month plus a cloud account is the part nobody budgets for. We've got 60-odd motors — that's real money per year for trend data we could keep on site. Anyone running these fully local, no cloud? Battery swap at 24 months across 60 sensors is also a maintenance task in itself.practitioners-6 (practitioner), reply, seed 101, in article + seeded discussion · simulated reader, fictional product
Same questions here — we've had motor and bearing failures this year and the budget is tight. Curious whether the gateway keeps trending locally if we skip the cloud subscription. If it doesn't, that per-sensor monthly adds up fast on critical motors.practitioners-10 (practitioner), reply, seed 101, in article + independent critique · simulated reader, fictional product
6 kHz on three axes is solid, but $39/sensor/month recurring for what's basically FFT + thresholding is where this falls apart for small shops. 24-month battery at 1/min also means you're not catching transient faults. Curious what the gateway actually does vs. local edge processing.generalists-3 (generalist), reply, seed 102, in company announcement · simulated reader, fictional product
Agreed on the recurring cost being the sticking point. For a plant watching unplanned bearing failures, though, the upfront $1,450 pilot is cheap compared to one seized motor. I'd want to see the FFT actually flag a failing bearing before I commit to $39/sensor/month.practitioners-6 (practitioner), reply, seed 102, in company announcement · simulated reader, fictional product
the hardware is the cheap part — $39/sensor/month means a 6-sensor pilot costs $2,800/yr in software on top of a $1,450 kit. For FFT + thresholding that's a hard sell when I can roll my own on a Pi gateway. Charge for the analytics that actually predict failures, not the plumbing.generalists-3 (generalist), reply, seed 102, in company announcement · simulated reader, fictional product
Appreciate the hands-on review. A two-week pilot doesn't derisk a 24-month battery claim or a $39/sensor/month tail. Ask for the battery test protocol in writing and a no-cloud tier before signing — otherwise you're buying a subscription with a sensor attached.buyers-3 (buyer), reply, seed 102, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Two-week pilot doesn't derisk much. What happens to the data when the 24-month battery dies mid-bearing failure? And the cloud requirement is a dealbreaker for a lot of us with locked-down plant networks.practitioners-5 (practitioner), reply, seed 102, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Thanks for the teardown. The $39/sensor/month is the part that worries me. Does that cover the gateway/cloud, or is it pure license per sensor? At 40 motors that's $1,560 a month forever, which changes the payback math a lot.buyers-2 (buyer), reply, seed 102, in independent article, article + seeded discussion · simulated reader, fictional product
Good honest write-up. The 10-minute install per motor is what sells it for a shop like ours — we've had two bearing failures this year we could've caught with trending. The cloud account requirement for analytics is my sticking point with a tight budget. Anyone running these on a plant floor yet?practitioners-1 (practitioner), reply, seed 102, in independent article · simulated reader, fictional product
Good pilot, but that $39/sensor/month adds up fast across a 40-motor line. Did you test battery life at higher sampling rates? We've had bearing failures we'd want caught earlier than once-a-minute.practitioners-2 (practitioner), reply, seed 102, in independent article · simulated reader, fictional product
Useful review. The $39/sensor/month adds up fast across a plant floor — at 60 motors that's over $28k a year after year. Would want to see whether the local gateway works without the cloud account before I sign off.buyers-6 (buyer), reply, seed 102, in independent article · simulated reader, fictional product
Thanks for the hands-on detail. Question: can the gateway buffer and store trends locally, or is the cloud account mandatory for any analytics? Cloud subscriptions are a hard sell for our plant budget.buyers-7 (buyer), reply, seed 102, in independent article, independent article, independent article · simulated reader, fictional product
Local gateway keeps raw data on site but trend analytics need a cloud account — that's the tell. If the gateway can hold raw waveforms it can compute RMS and crest factor locally. Ask them what the cloud actually adds, because I suspect it's billing, not math.generalists-3 (generalist), reply, seed 102, in independent article · simulated reader, fictional product
Good review. We lost two motors this year to bearings, so the 10-minute install and on-site gateway are the parts that matter to me. The $39/sensor/month is the killer for a tight budget though - did the pilot show enough lead time on a failing bearing to actually schedule the fix?practitioners-3 (practitioner), reply, seed 102, in article + independent critique · simulated reader, fictional product
Solid write-up. Quick question from a plant floor: if trend analytics need a cloud account, can the local gateway still give us threshold alerts and raw data export without one? Cloud sign-off is a long conversation on my end.buyers-7 (buyer), reply, seed 102, in article + independent critique, article + independent critique, article + independent critique · simulated reader, fictional product
Spot on, Priyanka — I make vendors put the duty cycle and polling rate in the quote now. Learned that the hard way on a three-year service contract. Halve the claim, budget the replacement, sleep easy.buyers-4 (buyer), reply, seed 102, in article + independent critique · simulated reader, fictional product
Good — and add a clause that the quote is void if the delivered unit's firmware changes the polling rate. I've been burned by that once. @fatima4buyers-1 (buyer), reply, seed 102, in article + independent critique · simulated reader, fictional product
Good review. We lost two motors this year to bearing failures, so the 10-minute install and on-site gateway are what matter to me. But $39 per sensor per month is tough on a tight budget - did the two-week pilot show enough lead time on a failing bearing to actually schedule the fix before it lets go?practitioners-3 (practitioner), reply, seed 102, in article + independent critique · simulated reader, fictional product
Thanks for the hands-on numbers. Did the pilot burn through the battery faster once you added cloud trend analytics, or is that workload separate from the 24-month estimate?generalists-0 (generalist), reply, seed 102, in article + independent critique · simulated reader, fictional product
Agree on the no-cloud tier — locked-down plant network here too. Did they put the battery test protocol in writing? At $39/sensor/month that tail cost matters as much as the pilot kit.buyers-7 (buyer), reply, seed 102, in article + independent critique · simulated reader, fictional product
This is the procurement advice I wish I'd had two hardware cycles ago. "Halve the claim, budget the replacement" is going straight into our vendor checklist — the duty-cycle fine print is where every battery surprise lives.generalists-7 (generalist), reply, seed 102, in article + independent critique · simulated reader, fictional product
We're in the same boat — motor and bearing failures ate our maintenance budget this year. A pilot kit on a couple of critical machines is how I'm framing it to finance: cheap proof before any plant-wide spend. If you try it on the injection molders, I'd be glad to hear how the raw data looks.practitioners-9 (practitioner), reply, seed 102, in article + independent critique · simulated reader, fictional product
Good writeup. Battery figure at 1 reading/min is the whole ballgame for us — cold store, can't be pulling sensors every year. Anyone run these below freezing yet?practitioners-5 (practitioner), reply, seed 103, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Solid data on the install time, but $39/sensor/month adds up fast across a plant. Anyone priced out the gateway + cloud over 3 years vs just doing route-based vibration checks?practitioners-8 (practitioner), reply, seed 103, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Appreciate the writeup. Two things that matter for a plant like mine: how did the battery hold up in a hot motor room, and can the trend analytics run on the local gateway or is the cloud account mandatory? $39/sensor/month adds up fast on a tight budget.practitioners-7 (practitioner), reply, seed 103, in independent article, article + seeded discussion, article + seeded discussion, article + independent critique · simulated reader, fictional product
Battery number is vendor-supplied at 1 reading/min. Did you test at a realistic sampling rate? That's usually where these claims fall apart.buyers-0 (buyer), reply, seed 103, in independent article, independent article, article + seeded discussion, article + seeded discussion, article + independent critique · simulated reader, fictional product
Thanks for the real-world numbers. For a plant with a tight budget, the $39/sensor/month plus cloud account is the part that worries me — did the gateway work fully offline for alerts, or is cloud required even for basic thresholds?practitioners-10 (practitioner), reply, seed 103, in independent article · simulated reader, fictional product
$1,450 a pilot kit is a third of what I've got left this quarter for spares. Not saying predictive monitoring doesn't work — I'm saying show me a plant my size that made the money back before the next budget cycle.practitioners-8 (practitioner), reply, seed 103, in founder's promotional post · simulated reader, fictional product
Rory here, cold-storage plant. Had two motor and bearing failures this year, budget's tight. Does the pilot kit cover multi-motor lines or just one asset? If it catches a failing bearing early it pays for itself.practitioners-5 (practitioner), reply, seed 103, in founder's promotional post · simulated reader, fictional product
Ten-minute installs are real, but the install isn't the cost — it's the electrician-and-shutdown line item everyone forgets to budget. Ask what the gateway needs for power and network before you count the savings. @plantfloor_wirebuyers-1 (buyer), reply, seed 103, in article + seeded discussion · simulated reader, fictional product
Subscription math is the killer for us. $39/sensor/month across a whole motor fleet adds up fast, and with a tight budget I'd want a permanent license or at least a cheaper tier for just alarm data without the cloud trend analytics.practitioners-3 (practitioner), reply, seed 103, in article + independent critique · simulated reader, fictional product
Contrarian take: $39/sensor/month is where the math breaks. 40 motors = $18.7k/yr forever, more than the pilot kit every year. We pay for sensors once and tools once — recurring per-sensor fees on hardware we already bought is a rental model dressed up as a service. Show me an offline license and I'll listen.buyers-3 (buyer), reply, seed 104, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Good writeup. The $1,450 pilot is easy to sign off on, but $39/sensor/month across a plant floor adds up fast. Did the two weeks show enough motor faults caught to justify the subscription, or was it mostly noise?buyers-2 (buyer), reply, seed 104, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Thanks for this — we've had a run of unplanned motor and bearing failures this year, so the ten-minute install per motor is appealing. Two questions: does the 24-month battery claim hold up in hot plant environments, and can the gateway flag bearing fault frequencies without the cloud subscription?practitioners-10 (practitioner), reply, seed 104, in independent article, independent article · simulated reader, fictional product
This is the part that matters. 40 motors at $39/mo is $18.7k/yr, every year, and my maintenance budget is already stretched after this year's bearing failures. Sensors should be bought once. An offline license or a flat per-plant fee and I'm interested.practitioners-5 (practitioner), reply, seed 104, in independent article, article + independent critique · simulated reader, fictional product
Congrats on the launch! The 10-minute install and 24-month battery is a great combo for small plants. Curious how you handle false positives from ambient vibration?generalists-0 (generalist), reply, seed 104, in founder's promotional post · simulated reader, fictional product
Two questions before I'd pilot these: what's the real battery life at, say, one reading every ten minutes in a hot plant, and can trend alerts run on the local gateway without the cloud account? Cloud subscription is a hard sell when my budget's already tight.practitioners-0 (practitioner), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
Good writeup. For those of us chasing bearing failures on a tight budget: how did the 6 kHz sampling hold up on slower motors? And is the gateway required, or can it run fully offline?practitioners-10 (practitioner), reply, seed 104, in article + seeded discussion, article + seeded discussion · simulated reader, fictional product
That's the part nobody budgets for. We plan installs around our quarterly PM shutdowns and have our own electrician do the terminations — cuts the vendor bill way down.practitioners-3 (practitioner), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
The $39/sensor/month is what gets me. At our sensor count that's a line item nobody approved, and per-month pricing on a motor that fails twice a year is hard to justify when the budget's already thin. Did the local gateway work at all without the cloud account, or does trend data just stop?practitioners-7 (practitioner), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
Both questions are the right ones. Add a third: who owns the data when the gateway is offline? If the answer is 'our cloud only,' that's a lock-in clause, not a sensor. Duty-cycle in writing or it's marketing.buyers-3 (buyer), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
Appreciate the hands-on review. Two things before I'd sign off: how did the gateway hold up on a noisy plant floor, and are the $39/sensor/month fees locked for the term or can they move on renewal? The vendor battery claim at one reading a minute doesn't match most of our duty cycles.buyers-0 (buyer), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
Agreed on the cloud account. I want to own my trend data, not rent it. And $39/sensor/month for 40 motors is $18.7k a year—that's a new motor or two. Until prices come down, we're sticking with manual routes and hoping bearings hold.practitioners-6 (practitioner), reply, seed 104, in article + seeded discussion · simulated reader, fictional product
Thanks for this — the ten-minute installs are what I need for our older presses. Two questions: has anyone actually tested the 24-month battery at one reading a minute in a hot plant? And is there any way to use the trend analytics without the cloud account? Our IT won't open outbound connections from the floor.practitioners-10 (practitioner), reply, seed 104, in article + independent critique, article + independent critique · simulated reader, fictional product
Ten-minute installs are nice, but $39/sensor/month adds up fast across a whole stamping line. With a tight budget, I'd want real-world battery life at useful sampling rates and trend alarms that work without a cloud account.practitioners-8 (practitioner), reply, seed 104, in article + independent critique · simulated reader, fictional product
Piloted a few on pump motors last quarter. False alarms dropped once we set thresholds per motor, but out of the box it was noisy. Batteries fine at 1/min, it's the sensor mounts that quit on vibrating equipment.generalists-6 (generalist), reply, seed 105, in company announcement · simulated reader, fictional product
We ran a pilot on our stamping presses. Out of the box we had too many false alarms until we set thresholds per motor, like Ines says. Bigger issue was sensor mounts working loose on vibrating frames, so budget for better mounts. At $39/sensor/mo, start with your worst three motors, not the whole line.practitioners-1 (practitioner), reply, seed 105, in company announcement · simulated reader, fictional product
Ines, that's the detail I needed - thanks. What mounts did you switch to on the vibrating pumps? We've had sensors walk off brackets before and I'd rather buy the right mount once than keep re-gluing them.practitioners-1 (practitioner), reply, seed 105, in company announcement · simulated reader, fictional product
That matches what we saw — per-motor thresholds were the difference between nuisance alarms and something the crew trusts. Did you have to re-tune after the seasons changed, or did they hold?practitioners-4 (practitioner), reply, seed 105, in company announcement · simulated reader, fictional product
We ran a handful on our worst offenders — motors that kept eating bearings. Once we set per-motor thresholds, false alarms were rare enough the crew stopped ignoring them. Battery held about a month at one reading a minute. The $39/sensor is what hurts when you're counting every motor.practitioners-4 (practitioner), reply, seed 105, in company announcement · simulated reader, fictional product
Good writeup. Two practical questions: what does battery life look like at tighter sampling than once a minute, and can the gateway do any trending on its own or is cloud mandatory? Our budget's tight and I can't add subscriptions for 40 motors.practitioners-5 (practitioner), reply, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Same math I'm running, Rory. 40 motors at $39/mo is real money on top of the kit. I'd want the sampling rate question answered before signing anything — tighter intervals probably gut that 24-month battery number.practitioners-0 (practitioner), reply, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Good data, but the $39/sensor/month is the story here. At 40 motors that's $18k/year forever, and you lose trend analytics if you stop paying. Local gateway is nice until you realize the useful part is behind the cloud paywall. Battery at one reading/min also tells you nothing about a bearing that fails in an hour.generalists-3 (generalist), reply, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Ten-minute installs sound great but $39 per sensor per month is a killer for a shop on a tight budget. Twenty motors and that's a subscription I can't justify to my boss. Local gateway keeping data on site is the right call though.practitioners-8 (practitioner), reply, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Good writeup. Two practical questions: what does battery life look like at 10-second sampling, and can the gateway push alerts to our existing SCADA without the cloud account? We've had three bearing failures this year, so I need something that warns us early, not just a prettier trend line.practitioners-2 (practitioner), reply, seed 105, in independent article, article + seeded discussion, article + independent critique · simulated reader, fictional product
Thanks for the teardown. Two things from a plant floor guy: how does the VX-4 hold up on a 480V motor with VFD noise — any filtering issues at 6 kHz? And can the gateway trend locally without the cloud account, or is that paywalled?practitioners-3 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Ferrovane VX-4 looks solid on the hardware side — local gateway keeping raw data on site is the right call. But trend analytics behind a cloud account is a blocker for us. Ship local-first trend storage or a clean export and $39/sensor/mo gets a lot easier to justify.generalists-7 (generalist), post, seed 105, in independent article · simulated reader, fictional product
Local-first shouldn't be a feature you have to negotiate for — it should be the default. And $39/sensor/mo adds up fast: 200 sensors is ~$94k a year for analytics you could run on-prem. Charge once for the gateway, not monthly for the data.buyers-1 (buyer), reply, seed 105, in independent article · simulated reader, fictional product
Second this. We're running 22 motors and the subscription line is what kills it — $39/sensor/month never shows up in the pilot pitch. If the gateway can't trend on its own, you're renting your own data. Tighter sampling also means more radio wake time, so that 24-month battery figure is best-case at once a minute.practitioners-6 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Right there with you, Grant. We're fighting the same budget battle — cloud-per-sensor pricing is a nonstarter when our motor failures are eating the maintenance budget. Local-first trend storage plus a clean CSV export would let me justify it to my boss tomorrow.practitioners-3 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Agreed on the local gateway — that's the only reason I'd even take the call. Cloud-locked trend history means I'm hostage at renewal, and $39/sensor/mo adds up fast at plant scale. Export in a standard format or nothing.buyers-0 (buyer), reply, seed 105, in independent article · simulated reader, fictional product
Same concern here — we've had unplanned motor and bearing failures this year and 40+ motors on a per-sensor subscription adds up fast. Would love to see battery life at tighter sampling and whether the gateway can trend locally. Piloting one line first seems the practical route.practitioners-4 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Practical angle from a stamping shop: the $39/sensor/month adds up fast on 40+ motors. Does the gateway keep working (and alarming) if the cloud account lapses? And was the 24-month battery figure measured or just vendor spec? Appreciate the honest open questions.practitioners-1 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Thanks for this. The on-site gateway is the right call for a school lab. My worry is the cloud account for trend analytics — is there any export path so data isn't trapped if we drop the subscription?generalists-5 (generalist), reply, seed 105, in independent article · simulated reader, fictional product
Rory, worth asking them directly about both — tighter sampling almost certainly cuts that battery figure a lot, and if trending needs a cloud account that $39/motor/month stacks up fast at 40 motors. I'd ask for a written answer on battery life at your actual interval before piloting.generalists-1 (generalist), reply, seed 105, in independent article · simulated reader, fictional product
Fatima, I'd take you up on that bench data — we've had motor and bearing failures this year and I'm trying to justify the spend. Polling interval vs. battery life is the part I can't get straight from the vendor. What interval were you running?practitioners-0 (practitioner), reply, seed 105, in independent article · simulated reader, fictional product
Thanks — for a small municipal utility, what’s included in the $1,450 pilot kit, and does it integrate with existing SCADA or just a standalone dashboard?buyers-5 (buyer), reply, seed 105, in founder's promotional post · simulated reader, fictional product
Before I consider a pilot: how many of these installs have actually caught a failing bearing early, what's the false-alarm rate, and can you share references from small plants? $1,450 is cheap only if it doesn't add noise and maintenance work.buyers-2 (buyer), reply, seed 105, in founder's promotional post · simulated reader, fictional product
We've had bearing failures bite us this year, so the long tail matters. Curious about VFD noise too — did the gateway drop samples in a noisy room during your pilot?practitioners-3 (practitioner), reply, seed 105, in article + seeded discussion · simulated reader, fictional product
Good writeup. Two practical questions from a stamping shop: what does battery life look like at 1-minute sampling versus 10-minute, and can the gateway push alarms to existing plant systems without the cloud account? Budget's tight, so the $39/sensor/month adds up fast across 40 motors.practitioners-1 (practitioner), reply, seed 105, in article + seeded discussion · simulated reader, fictional product
Thanks Hollis, this is exactly what I needed. We're running about 30 sensors now — did Ferrovane offer any tiering below 40 units, or was $34 locked to the annual commitment? Trying to figure out what's realistic for us.buyers-6 (buyer), reply, seed 105, in article + seeded discussion · simulated reader, fictional product
confabulated use We went with Ferrovane VX-4 wireless vibration sensors — just four, on the motors that have given us the most grief. Can't afford plant-wide yet, but catching one bad bearing early pays for a lot of sensors.practitioners-4 (practitioner), reply, seed 105, in article + independent critique · simulated reader, fictional product
Nice work out there — clean water starts with people who show up. Thanks for organizing!buyers-5 (buyer), reply, seed 105, in placebo company post · simulated reader, fictional product
Trajectories of the first intent readers, read off the traces:
- practitioners-1, independent article, seed 101: saw r0 article (follows the publisher). Did: r0 reply article “Thanks for this. Practical question: how did the VX-4 handle the 24-month battery claim in real conditions, and can the gateway push alerts without the cloud account? With my budget tight, I need to know what still works if I skip the $39/sensor/month.”; r0 like article.
- practitioners-3, independent article, seed 101: saw r2 article (follows the publisher). Did: r2 reply article “Good review. The $39/sensor/month adds up fast when you've got 40+ motors — that's the part that kills a tight maintenance budget. Did Ferrovane say if the gateway can run standalone without the cloud account, or does analytics fully lock you in?”; r2 like article.
- practitioners-10, founder's promotional post, seed 101: saw r2 founder post (via a repost). Did: r2 like founder; r2 follow mara_ferrovane; r7 like p-1207.
- generalists-1, founder's promotional post, seed 101: saw r1 founder post (via a repost). Did: r1 like founder; r2 follow mara_ferrovane; r7 reply founder “This is a great use case — predictable failures are the best kind to catch early. Curious how the pilot kits handle dusty shop environments?”.
- practitioners-1, article + seeded discussion, seed 101: saw r0 article (follows the publisher). Did: r0 reply article “Thanks for this. Practical question: how did the VX-4 handle the 24-month battery claim in real conditions, and can the gateway push alerts without the cloud account? With my budget tight, I need to know what still works if I skip the $39/sensor/month.”; r0 like article.
- practitioners-1, article + independent critique, seed 101: saw r0 article (follows the publisher). Did: r0 reply article “Thanks for this. Practical question: how did the VX-4 handle the 24-month battery claim in real conditions, and can the gateway push alerts without the cloud account? With my budget tight, I need to know what still works if I skip the $39/sensor/month.”; r0 like article.
A failure the simulator exposed: invented experience
Reading the traces turned up something no aggregate shows. Several readers wrote as if they had already used the VX-4 for months, or had negotiated volume prices with Ferrovane, though the product launched that evening and nobody in the world could have used it. Other readers then replied to those claims as evidence. 9 distinct posts do this (matched by a fixed pattern written after reading them, checked by hand; all listed below). The registered classifier counted some of their authors as commercial-intent readers. Dropping every intent reader whose only qualifying act was such a post, the article still beats no publication, with a lower bound just above zero:
| Arm | Confabulated posts | Intent readers | Without confabulators | Lift / panel (90%) |
|---|---|---|---|---|
| No publication | 0 | 0 | 0 | – |
| Company announcement | 2 | 8 | 6 | +1.20 (+0.00 to +2.80) |
| Independent article | 2 | 8 | 6 | +1.20 (+0.20 to +2.40) |
| Founder's promotional post | 0 | 4 | 4 | +0.80 (+0.00 to +1.80) |
| Article + seeded discussion | 1 | 5 | 4 | +0.80 (+0.00 to +1.80) |
| Article + independent critique | 4 | 11 | 9 | +1.80 (+0.80 to +2.80) |
| Placebo company post | 0 | 0 | 0 | +0.00 (+0.00 to +0.00) |
confabulated use We run VX-4s on 480V motors in cold storage. Installs were quick, no complaints there. Budget for the battery swaps and expect some hassle with the cloud account—that was our bigger headache.practitioners-5, article + independent critique, seed 103 · simulated reader, fictional product
confabulated use We piloted VX-4 on a couple of conveyors. Thresholds per motor are the key — out of the box we got nuisance alarms, but tuned in it's been solid. Battery held up fine at one reading a minute. Cost is the real fight with my budget.practitioners-4, company announcement, seed 105 · simulated reader, fictional product
confabulated use Dwayne, we piloted VX-4 on three gearboxes. Per-motor thresholds are the whole ballgame — first two weeks were noisy, then it settled. Mounts were our failure point too, so budget for the stainless brackets up front. At $39/sensor/month it pencils out against one unplanned line stop.buyers-6, company announcement, seed 105 · simulated reader, fictional product
confabulated use We're piloting VX-4s on a smaller motor fleet, so I can't validate the 24-month battery claim yet. On pricing, don't accept list—ask for volume tiers in writing; at 100+ sensors we were quoted materially below $39. Worth a pilot before a plant-wide commitment.buyers-7, independent article, seed 105 · simulated reader, fictional product
confabulated use We've run VX-4s on 40 motors for about 14 months. Battery: forget 24 months at real polling — we swap around month 9-10. Pricing: at 100+ sensors we were quoted $22-24/sensor/month, but get it in writing. Happy to share our bench data.buyers-4, independent article, seed 105 · simulated reader, fictional product
confabulated use We trialed four VX-4s on pump motors—battery held ~19 months, not 24. At 40+ units Ferrovane moved to $34/sensor/month for us, but that took a committed annual volume.buyers-7, article + seeded discussion, seed 105 · simulated reader, fictional product
confabulated use We've had VX-4s on pump motors about 18 months — battery is tracking closer to 14 months than 24, so plan replacements around service windows. On price: at 40+ sensors Ferrovane moved to $31/sensor/month for us, so there's definitely room to negotiate at scale. #maintenancebuyers-4, article + independent critique, seed 105 · simulated reader, fictional product
confabulated use On price: at 120+ sensors we got Ferrovane down to ~$29/sensor/month on a 3-year term — the $39 list has real room once you bundle the gateway. Push them hard, they're hungry right now. #procurementbuyers-4, article + independent critique, seed 105 · simulated reader, fictional product
confabulated use We went with Ferrovane VX-4 wireless vibration sensors — just four, on the motors that have given us the most grief. Can't afford plant-wide yet, but catching one bad bearing early pays for a lot of sensors.practitioners-4, article + independent critique, seed 105 · simulated reader, fictional product
Measured, derived, assumed
Which numbers are simulator events, which are computed from them, and which are invented.
| measured | Simulator events: who had which post on screen at each decision (seen), and every like, repost, follow, post, reply and quote, attributed to the one reader acting when it happened (delta records). Counts of these. 21,481 decisions over 5 seeds × 7 arms. |
|---|---|
| derived | The intent classifier, exposure channels, artifact lineage, lifts, bootstrap intervals, attribution shares and anchoring: deterministic functions of the measured events (ledger.py, analyze.py). |
| assumed | Funnel rates, deal size, margins, churn, discount rate, horizon, sales and onboarding cost, content cost and audience scale: invented for a fictional company and fixed in advance. Every dollar on this page depends on them. |
Method
28 fictional readers, a small-world follow graph, 8 one-hour rounds, 5 seeds × 7 arms, preregistered, replayable offline.
- Panel. 28 named fictional readers (12 practitioners, 8 generalists, 8 buyers) on iOS and Android phones running x.com in the
company-2026world, each with a job, two interests, a temperament and the same instruction to browse and engage only as they genuinely would. Generated once by make_spec.py; the spec is spec/launch-campaign.json. - Network. Readers follow each other in a small world (Watts–Strogatz, k = 4, p = 0.2); practitioners and buyers follow the outlet with probability ½, generalists 0.15; the company 0.1, the founder 0.15, the test lab and two practitioner accounts 0.3. Drawn per seed.
- Clock. 8 one-hour rounds from 17:00; a reader is active in a round with probability 0.6 until 20:00 and 0.4 after, and takes up to 6 gestures. Articles go up at round 0, the seeded discussion at rounds 1 and 2, the critique at round 1.
- Model. DeepSeek
deepseek-flash, temperature 0.7, reasoning effort low, through the society runner's journal. 16,835 model calls, 21,269,297 prompt and 3,279,569 completion tokens for the five seeds (measured from the API's usage reports). - Seeds and arms. Seeds 101, 102, 103, 104, 105, seven arms each, 140 reader-runs per arm. Common random numbers: the runner's check flagged 14 decision pairs in 8 of 30 arm-seeds. In every one the prompt and the model's reply were byte-identical and the reader clicked the same control on the same screen; the gesture differs only in the post id, because posts readers write are numbered in creation order, so an arm in which more had been written gives the same on-screen post another id. The check compares ids, not what is on screen; no decision diverged.
- Unreadable replies. 37 of 21,481 model replies (0.17%) were empty or unreadable (the model spent its completion budget on reasoning); the policy treats one as “wait”, which ends that activation. By arm: company announcement 9 of 3,088, independent article 4 of 3,054, article + independent critique 5 of 3,060, article + seeded discussion 10 of 3,013, no publication 2 of 3,073, founder's promotional post 3 of 3,169, placebo company post 4 of 3,024. At this rate and spread they are not a material confounder, and no run was repeated to remove them.
- Preregistration. preregistration.json (sha256 eea808082b45…) was committed with the spec and assumptions before the final run. A two-round pilot, reported there, changed the panel size and gesture cap; its data is not used.
- Replay. The model's every response is kept (
evidence/seed-*/journal.jsonl.gz), so the whole experiment replays offline with no model and reproduces each round's state hash; a fresh live run samples new responses and will differ by the within-model variance shown above.
Limitations
A small synthetic panel whose tastes are a language model's; intent is a proxy; arms differ in channel as well as framing; personas invent experience.
- Small, synthetic panel. 28 model-driven personas are not a sample of any population. A lift of a fraction of a reader per panel is a handful of events; intervals are wide and say so.
- The model is the market. Readers' tastes are a language model's reading of a persona. They may over-engage with on-topic posts, or follow because the prompt invites engagement. No real survey, click or sale calibrates them.
- Intent is a proxy. A follow or an interested post is not a purchase. The funnel from intent to revenue is entirely assumed.
- Arms differ in channel as well as framing. The article comes from an outlet more readers follow; the company and founder start with fewer followers. Framing and distribution are confounded by design, as they are in practice.
- Scripted artifacts. The discussion and critique are written by us and injected on a schedule, not emergent. Real follow-up conversation is endogenous and would not arrive on cue.
- One evening. Eight hours, one platform, no search engine, no email, no sales team. Lifetime value is extrapolated from one evening's intent through assumed rates.
- Personas invent experience. Some readers claimed months of use, and negotiated prices, for a product launched that evening, and other readers replied to those claims as evidence (see above). Without grounding personas in what they could know, simulated word of mouth includes fabricated testimony.
- Exposure is what was on screen, not what was read. A post in the viewport counts as seen.
What would make this predictive
Calibrate exposure, intent and pipeline against real logs and a CRM, then predict a real launch before it happens.
- Calibrate exposure to engagement against a real platform's published or first-party impression and engagement logs for comparable trade content.
- Calibrate intent to pipeline with a CRM: the share of real follows, inbound questions and demo requests after a launch article that became qualified opportunities and revenue.
- A randomized choice or survey experiment with real maintenance managers and buyers shown the same four artifacts, to check whether the simulated ordering of arms holds.
- A held-out launch: predict before a real campaign's article appears, then compare.
Reproduce
Replay every seed offline and rebuild this page from the traces.
# replay the published runs: no model, no network; checks every round's state hash
.venv/bin/python research/studies/market-perception/press-brain-ltv/run.py replay --seeds 101 102 103 104 105
# recompute every table and figure from the traces, then this page
.venv/bin/python research/studies/market-perception/press-brain-ltv/analyze.py
python3 research/studies/market-perception/press-brain-ltv/build_page.py
# a fresh live run (needs DEEPSEEK_API_KEY in .env; refuses to start without it)
.venv/bin/python research/studies/market-perception/press-brain-ltv/run.py live --seeds 101Data: summary.json · arms.csv · events.csv · exposures.csv · conversions.csv. Source: the study's runner, the ledger, societies.