# PostHog synthetic prompt reset when real user language disagrees

> Throw away the prompt set and start over when synthetic tracking stops matching how real customers actually describe the product.

- Canonical HTML: https://growth.iangoh.com/growth-ideas/posthog-synthetic-prompt-reset-when-real-user-language-disagrees/
- Source: [newsletter.posthog.com](https://newsletter.posthog.com/p/llms-are-picking-winners-heres-how)
- GrowthDex source hub: [PostHog: LLMs are picking winners. Here's how to become one.](/sources/posthog-llms-are-picking-winners-here-s-how-to-become-one-newsletter-pos/)
- Last checked: 2026-06-09T14:12:06.000Z
- Rarity: rare
- Budget: low
- Channels: AI visibility, Analytics, Positioning
- Stages: prompt reset, data quality, adaptability, positioning


## Why this can grow

New channels tempt teams to defend bad history because the data feels scarce and therefore sacred. PostHog took the harder but better route. Once real prompt submissions showed that the tracked prompt set had little overlap with what real users were typing, the team scrapped the old system and rebuilt it. That hurts in the short term, especially when months of trend lines disappear. But keeping a false baseline is worse. It teaches the wrong topics, the wrong comparisons, and the wrong positioning. In fast-changing answer engines, adaptability is not cleanup work. It is the work.

## Ian's take

From scaling consumer platforms across MENA and Southeast Asia, my default is to distrust growth work that only looks good in a slide. My bias is to treat this as a small market test first. Make the audience narrow, make the promise concrete, and let the first real response decide whether it deserves more work. I would run it small enough to learn quickly, then only scale the parts that real users repeat, save, reply to, or buy from. For this tactic, I would watch one clear growth signal before putting more time or budget behind it.

## Action plan

1. Define one narrow startup segment where posthog synthetic prompt reset when real user language disagrees can create a measurable lift.
2. Turn the tactic into one offer, page, campaign, or workflow for the AI visibility and Analytics channel.
3. Use the evidence from newsletter.posthog.com to set the first version of the message, format, and audience.
4. Launch a small test for 7 to 14 days with one success metric: one measurable growth signal.
5. Review the result, keep the winning message, remove weak variants, and turn the learning into a repeatable growth playbook.

## Source-backed example

PostHog says it threw away roughly six months of AI-generated prompt history after realizing those prompts barely overlapped with the real user prompts collected through onboarding.

## Adjacent tactics in the same lane

- [PostHog AEO reporting quilt before single-source certainty](/growth-ideas/posthog-aeo-reporting-quilt-before-single-source-certainty/) - same source, 2 shared channels
- [PostHog citable content chunks before monolithic SEO pages](/growth-ideas/posthog-citable-content-chunks-before-monolithic-seo-pages/) - same source, 1 shared channel
- [PostHog private-mode three-prompt benchmark before AEO tool sprawl](/growth-ideas/posthog-private-mode-three-prompt-benchmark-before-aeo-tool-sprawl/) - same source, 1 shared channel
- [PostHog owned-surface cleanup before off-domain AEO sidequests](/growth-ideas/posthog-owned-surface-cleanup-before-off-domain-aeo-sidequests/) - same source, 1 shared channel

## Read GrowthDex essays

Browse the plain-English essay index at [GrowthDex Blog](/blog/).

## Related GrowthDex essays

- [The answer should be easy to quote before you chase the mention](/blog/the-answer-should-be-easy-to-quote-before-you-chase-the-mention/) - AI visibility, SEO, content strategy

## Advisory

If you want help turning this into a working growth system, Ian Goh offers advisory at https://iangoh.com/advisory.