An Amazon seller sits on more demand data than almost any supplier a retail buyer will meet this year. The problem is that it is expressed in a language the buyer does not speak. Rank, search volume, session conversion and advertising cost are precise terms about a marketplace, and a category manager at a regional chain has no shelf equivalent for any of them.
Translation is the job. Not dumping a dashboard into a slide, but converting the handful of numbers that carry meaning into the two or three sentences a buyer can act on. This article covers which numbers survive translation, which do not, and where the third-party estimates in your analytics stack become a liability rather than an asset.
Use your own reports, not a third-party estimate
Start with a rule that sounds obvious and is broken constantly. When a number is about your own product, take it from Seller Central, not from an analytics tool that models it.
Analytics products estimate sales for listings they do not own, which is a useful capability aimed at competitors. For your own items, the estimate is strictly worse than the truth, which you can export in a minute. Quoting an estimate about yourself is an unforced error: it will be slightly wrong, and one day a buyer with category data will notice.
So the workflow is to use the analytics stack for the category and the competitive landscape, and your own reports for anything about you. What to pull from those reports is set out in Amazon brand analytics, and the related boundary between marketplace research and distribution research is covered in product research tools and wholesale.
What translates, and into what
| Amazon metric | What a buyer hears | Say it as |
|---|---|---|
| Units sold, trailing year | Proven demand | Units per month, plus region |
| Review count and rating | Shopper trust | The count and the rating, plainly |
| Repeat purchase rate | Consumable, restocks | Share who bought again |
| Sales by state | Demand in their footprint | Rank of their states for you |
| Search volume for the category | Category interest | Skip unless it is large |
Four of those five are worth saying out loud in a first meeting. Units and region together are the strongest, because a buyer's first question about a new brand is whether anyone wants it, and their second is whether anyone near their stores wants it.
Repeat purchase rate is the underused one. For a consumable, it answers the buyer's real anxiety, which is whether the product sells once as a novelty or refills a slot every few weeks. A high repeat rate is the difference between a gift item and a shelf staple.
What does not translate
Rank. A best seller rank is a position in a system the buyer does not use. Saying you are ranked in the top few hundred of a subcategory means nothing to them, and worse, it invites the question of how large that subcategory is.
Advertising metrics. Cost per click, advertising cost of sales, impression share. These describe how you buy traffic on a marketplace. A shelf has no auction. Keep them internal, where they matter for the margin comparison in ACoS versus wholesale margin.
Session and conversion data. The conversion rate of a listing is a fact about your photography, your reviews and your price against other listings. It says nothing about whether the package sells from four feet away with no photography at all.
Keyword rankings. The single most Amazon-specific thing in your stack. There is no keyword on a shelf.
Estimated competitor sales. Directional at best, and dangerous in a room where somebody has the real number.
The principle underneath all five: anything that only exists because the marketplace has search and an auction does not survive the trip to a store.
The one page that carries the translation
Everything above condenses into one page you attach to a pitch. Six lines, all arithmetic, no adjectives.
- Units sold per month across the last twelve months, with the seasonal shape in one sentence.
- Rating and review count, as of a date.
- Repeat purchase rate if the product is consumable.
- The five states where your sales concentrate, and where the buyer's stores sit in that list.
- The price you sell at on the marketplace, and the wholesale price and suggested retail you are proposing.
- One line on what customers say most, drawn from reviews.
That last line does more than it looks. The recurring praise in your reviews is the product's actual selling proposition, discovered rather than invented, and the recurring complaint is the objection the buyer will raise anyway. Answering it first is what separates a supplier from a pitch.
Keep the page to one side of paper. A buyer forwards it internally, and long documents do not get forwarded. What else belongs in the pitch is covered in how to sell Amazon products in retail stores.
Where the analytics stack still earns its subscription
None of this means turn the tools off. Three jobs genuinely need them while you are building a wholesale channel.
Category mapping. Understanding how many competitors exist, how they cluster by price, and whether the category is consolidating. This is context for every conversation and it is not in your own reports.
Competitor tracking that points at the physical world. Identifying the brands that share your category, then going to find out which of them are already on shelves. A competitor already stocked in the chain you are pitching is a fact worth knowing before the meeting, and a competitor stocked nowhere is a hint that the category buys through a distributor rather than direct.
Watching the marketplace side while you are distracted. A wholesale push takes attention away from listings and advertising. Alerts on rank movement, listing changes, buy box loss and new sellers keep the channel that pays for everything from drifting while you are on the phone to buyers.
That third use is the one to keep running. The most common way a first wholesale push loses money is not the retail order, it is the marketplace revenue that softened while nobody was watching.
The numbers you will start needing instead
Once the retail channel exists, the metrics that matter are not in your Amazon stack at all, because they describe accounts rather than listings: accounts opened, revenue per account, days to first reorder, the share of accounts that reorder, and sell-through per store per week where the retailer shares it.
None of those has a dashboard you can buy into readily, and most brands assemble them from invoices and an account record. How to build that view is in analytics for the wholesale channel, and the layers of tooling underneath it in tools for wholesale.
The input that no analytics product gives you is the list of companies worth pursuing in the first place. Paste your listing into WholesalePilot and the preview shows which distributors and retailers plausibly stock products like yours, which is the starting point the translated one-pager is then written for.
Questions Amazon sellers ask
Should I show a buyer a screenshot from my analytics tool? No. Screenshots of dashboards read as unverified. A clean line of text with a number and a date reads as a supplier.
Is best seller rank ever worth mentioning? Only as shorthand with someone who sells on the marketplace themselves. With a store buyer, units and reviews say the same thing more usefully.
Can I use category data to argue for shelf space? Yes, carefully. The strongest version is a price tier or a product type the buyer's assortment is missing, evidenced by what the category contains.
Do buyers verify these numbers? Rarely at first, and thoroughly if the relationship grows. Assume everything you say will eventually be checked against your invoices.
What if my marketplace volume is small? Then lead with something else: the repeat rate, the rating, the regional concentration, or a sell-through number from an independent store. Small and specific beats large and vague.