Write Down What the Founder Knows
Early-stage stores often sell on the founder's replies. When she steps back, that judgment leaves with her. How to capture it before it fades.
If your store converted better in its first year than it does now, and you can’t quite say why, check one thing before you blame the market, the ads or the platform.
Did the person who used to answer the messages stop?
What was actually doing the selling
In an early-stage store, the founder usually answers pre-purchase messages herself. And her replies tend to look nothing like customer service:
“Honestly, take the smaller one. The large is heavy, and if you’re carrying it daily you’ll regret it.”
“That green photographs brighter than it is. In person it’s closer to sage.”
“I’d skip this for a gift unless you know her size. The scarf is safer.”
Specific. Opinionated. Based on knowledge that exists nowhere else in the business. And often recommending the cheaper option, or advising against a purchase entirely.
That last part matters most. A founder willing to say “don’t buy this” is a founder whose “buy this” gets believed.
It’s also why these conversations sell out of proportion to how few of them there are. The shopper who writes in before buying is usually one step from a decision and missing one fact she can’t get from the page. The founder supplies that fact, plus a verdict, and the decision gets made.
What happens when she steps back
The store grows. She hires someone for the inbox and moves on to the things that need her more: suppliers, stock, the next launch.
Support is measured on response time and satisfaction. Those reward fast, pleasant answers. They don’t reward knowing how a color really looks, and they actively discourage the reply that talks someone out of a sale.
So answers become quicker, friendlier and emptier: “Great question! Our sizing guide is linked here.”
Nobody did anything wrong. The founder’s knowledge simply never got transferred, and the new role is structurally a different job. She was selling. Support is resolving tickets.
The drop in conversion comes months later, gradually, during a period of growth, so it gets blamed on colder traffic and nobody connects it to the handover.
Spec versus judgment
When stores finally write this down, the usual result is a better FAQ. That misses the point, because what sold things was never the spec.
| The spec (already on the page) | The judgment (what actually sold it) |
|---|---|
| Large: 15 × 12 × 5.5 in | Heavy even empty. Commuters should take the medium. |
| Available in sage | Photographs brighter than it is. For bright, take the teal. |
| Sizes XS–XL | The cut is particular. Not a safe gift unless you know her size. |
The left column answers “what is it?” The right column answers “should I, and which one?”, which is the question the shopper actually wrote in with.
One test when writing: would she have said it out loud to a customer? If it reads like a product description, it’s the wrong kind of sentence.
Four steps
1. Check whether it’s happened. Compare conversion on sessions that included a pre-purchase conversation, before and after the founder stopped answering. If it fell and little else changed (traffic mix, pricing, catalog), you’ve found it.
2. Read her old replies next to your current saved replies. Give it an hour, side by side. The difference is usually obvious within minutes: hers name a product and give a reason, the new ones point to a page.
3. Mine the sent folder while she still remembers. The messages persist. Her memory of why she said each thing fades. Sit with her for an afternoon and turn replies into sentences you can reuse. Go product by product, and for each one ask three questions: who shouldn’t buy this, what does the photo get wrong, and what does she recommend instead when it’s the wrong fit.
4. Keep the answers that discourage. The instinct is to polish everything into selling points. Don’t. “Not a safe gift unless you know her size” is what makes every other recommendation believable.
What to write, and in what shape
Keep each entry short and attached to a product, not filed under a topic. A shopper asking about the large tote doesn’t need your returns philosophy. She needs the one sentence about weight.
A workable shape is three lines per product:
- Who it’s for, and who it isn’t. “Great for weekends. If you commute on foot, take the medium.”
- What the page gets wrong. Color that shifts under a camera, fabric that looks stiffer than it is, a size that runs small.
- The alternative. What she’d point to instead, and why.
Write them in her voice, not a brand voice. If the sentence would sound odd coming from a person holding the product, rewrite it until it wouldn’t.
And revisit the list whenever a product changes. Judgment about last season’s cut is worse than no judgment at all.
This isn’t about the founder going back to the inbox
A store that only sells well while one person is answering messages is fragile. The goal is the opposite: let the store keep selling the way she sold, after she’s stopped.
Which means the written-down judgment has to reach the shopper at the moment she has the question, not sit in a shared document. A shopper who has to find the answer herself usually doesn’t; she leaves, or she buys the wrong size and sends it back. The same knowledge also makes a second visit easier, which is the argument in why the second order is the cheapest sale.
Where Immerss fits
Immerss puts an AI sales agent on your store that answers from your catalog and from what you’ve taught it about your products, including the honest “take the smaller one.” It replies instantly, 24/7, adds the right item to the cart and walks the shopper to checkout. When a question needs a person, one tap brings a real advisor onto live video who can show the product and add it to her cart on the call.
It also keeps the record going forward. Every conversation is visible to you, so what shoppers ask becomes knowledge the store keeps, instead of knowledge that lives in one person’s inbox.
It installs on Shopify, WooCommerce, Salesforce or any store by script, and it’s free to start. See how other stores use it in our customer stories, or get started.
Your store may not have lost anything it measures.
It lost the one thing it never had to measure, because she was always there doing it.
Questions
Why does conversion drop when the founder stops answering customer messages?
Because the founder was usually selling, not supporting. Her replies carried judgment that exists nowhere else in the business: which size to take, how a color looks in person, when not to buy. Support roles are measured on response time and satisfaction, which reward fast, pleasant answers rather than that judgment. The knowledge never gets handed over, and the drop shows up gradually months later, so it gets blamed on traffic.
How do I check whether the founder handover hurt my store's conversion?
Compare conversion on sessions that included a pre-purchase conversation, before and after the founder stopped answering. If it fell while traffic sources, pricing and catalog stayed roughly the same, the handover is the likely cause. Then read a sample of her old replies next to your current saved replies; the difference is usually obvious within minutes.
Isn't a good FAQ page enough to capture product knowledge?
No. An FAQ captures the spec: dimensions, materials, the size range. What sold was the judgment on top of it, such as that the large bag is heavy even empty or that a color photographs brighter than it is. A useful test when writing: would she have said this sentence out loud to a customer? If it reads like a product description, it is the wrong kind of sentence.
Should the captured answers ever talk a customer out of a purchase?
Yes, keep those. A founder who says "don't buy this for a gift unless you know her size" is a founder whose recommendations get believed. Polishing every answer into a selling point removes the honesty that made the rest of the advice credible.
