Most people prospecting ecommerce merchants read the wrong artifact. They open the homepage, skim the About page, glance at the Instagram grid, and form an opinion about the brand. The homepage is a pitch. It is written to make the business look like the business the founder wishes they were running. The catalog is different. The catalog is what the company actually does every day: what it buys, what it photographs, what it prices, what it ships. It cannot be aspirational, because every line of it cost somebody real work.
If you sell to merchants — conversion work, ads, email, SEO, fulfilment, photography, a SaaS tool — the catalog is the cheapest high-signal read available to you. ELDB exposes productCount on every store record and the items themselves at the catalog endpoint. What follows is how to read that data the way an operator would, instead of treating it as one more number in a filter.
Price bands are a proxy for the entire business model
Price tells you the acquisition maths the merchant lives inside. Under about €25 average order value, you are looking at a volume business: it needs paid social at scale or marketplace distribution, it makes very little per order, and it almost certainly cannot absorb a €3,000 monthly retainer unless the volume is genuinely large. Between roughly €40 and €120 sits the band where most agency offers work, because a modest lift in conversion or repeat rate pays for the engagement. Above €300 you are in considered-purchase territory: long decision windows, trust and content matter more than creative volume, and the merchant can cover your fee out of a handful of extra orders.
The arithmetic worth doing before you write the email is simple. How many incremental orders does your fee cost them? A €2,000 monthly retainer against a €35 order at 40% gross margin needs roughly 143 extra orders a month just to break even. The same retainer against a €280 order at 55% margin needs thirteen. That single ratio explains why an identical pitch lands with one merchant and bounces off another, and it is computable from the catalog in seconds.
Read the distribution, not the average. A catalog with a €19 median and three €400 items is a completely different animal from a flat €120 catalog. Bimodal pricing usually means a core product plus consumables or accessories, which means repeat purchase exists, which means retention and email are the live problem rather than acquisition. A long flat band with no entry product means every visitor faces the full price on the first visit, which is a merchandising conversation.
SKU depth: forty, four hundred, four thousand, forty thousand
These are four different companies, and productCount separates them for free. Around forty SKUs you have a founder-led brand selling its own product. The catalog is curated because one person curated it. Its problem is demand, not merchandising, and its search surface is tiny. Around four hundred, somebody has been hired. Photography and copy are now a recurring cost line, category structure starts to matter, and internal search becomes a real conversion lever. This is the band where merchandising and conversion work pays back fastest.
At four thousand SKUs you are usually looking at a distribution business reselling brands it does not own. Its moat is availability and price, its search surface is enormous, and that surface is almost always neglected: supplier copy on product pages, thin category pages, pagination handled badly. This is where technical SEO and feed management are worth real money, because the inventory of opportunity is already sitting on the server.
Past forty thousand, the catalog is a data problem. Nobody is hand-writing anything. The buyer is not the founder, it is an ecommerce manager with a product information system and a shopping feed that is silently rejecting a slice of the inventory every week. Sell systems, not services, and expect a procurement process. The same email that charms a forty-SKU founder reads as noise to this person.
Depth inside a category beats breadth across categories
There is a structural point that people who work on ecommerce search make repeatedly and that almost nobody applies to prospecting: a store cannot compete for a query whose results page is full of listing pages containing fifty items if the store itself has three items in that category. Category depth determines what a store is allowed to rank for. Two stores with an identical productCount of 340 can be structurally unequal — twelve categories of eight items is a weaker business than two categories of sixty.
This is also a diagnosis you can hand over for free, which is what makes it useful in outreach. Something like: you have 340 products spread across 22 collections, and 17 of those collections hold fewer than ten items, which means they cannot win a category query and are mostly absorbing crawl budget and confusing your internal search. That is a specific, checkable statement about their business, delivered by a stranger who did the work.
Cross it with organicKeywords, which ELDB carries on 80,593 stores. A deep catalog with a low keyword count is latent inventory. A shallow catalog with a high keyword count is a brand that has more demand than product, which is a very different conversation and often a better one.
Variants, options and the hidden operational load
Product count and variant count diverge wildly by vertical, and the gap is diagnostic. Apparel with sizes and colours can run forty products against six hundred sellable variants. That gap is where stockouts, size-related returns, and inventory headaches live. A store with a 1:15 product-to-variant ratio has an operations problem waiting to become a customer service problem.
If you sell inventory tooling, returns software, or anything operational, this ratio is your qualifying question and you can answer it before contact. If you sell creative work, the same ratio tells you how expensive a photography or content project will actually be, which stops you quoting a number you will regret.
What the catalog says about the constraint the operator is feeling
Catalog data carries traces of the person maintaining it. Titles that all begin with the brand name are one person’s convention, held consistently, which usually means a small and disciplined team. Titles carrying supplier boilerplate mean nobody has time to rewrite anything. Descriptions of wildly uneven length mean the catalog was built in bursts by different people. None of this is in a marketing brochure, and all of it is visible in the data.
From that you can usually name the real bottleneck: production, traffic, or conversion. A tight, well-maintained catalog with low traffic has a demand problem. A sprawling, half-finished catalog with strong traffic has a conversion and merchandising problem. Saying which one you think it is, in the first two sentences of an email, is worth more than any amount of flattery.
Cross-reading: catalog against traffic, keywords and platform
The catalog read gets sharper the moment you put it next to the other fields. Four thousand SKUs and three thousand monthly organic visits gives a keyword-per-product ratio near zero, which is a large, specific, quantified opportunity. Forty SKUs and ninety thousand visits is the mirror image: a brand with demand whose constraint is that it has nothing else to sell to an audience that already trusts it. ELDB has monthly traffic on 118,324 stores and organic keyword counts on 80,593, so both ratios are computable at list level rather than store by store.
Platform changes what the same catalog means. The distribution is lopsided — 58,549 WooCommerce stores, 50,002 Shopify, 16,501 PrestaShop, 15,477 JouwWeb, 4,510 Odoo — and each implies a different buyer. A 6,000-SKU WooCommerce store is a performance and infrastructure conversation. The same catalog on Shopify is a merchandising and app-stack conversation. On Odoo it is attached to an ERP, which means the decision-maker sits in operations and the sales cycle is longer than you want it to be.
From read to opening line
The read is worth nothing until it becomes a sentence the merchant recognises as true about their own business. That is the only test. “I noticed you have a lot of products” fails it. “You are running 2,100 SKUs across 31 collections, and only 14 of those collections have more than twenty items” passes it, because they can verify it in ten seconds and they have probably felt it without articulating it.
The catalog read also protects you from calls you should never take. A store with a €14 median price, thirty SKUs and four hundred monthly visits has no capacity to pay you, no matter how nice the founder is. Disqualifying early is not pessimism, it is what lets you spend real effort on the accounts that can actually convert. Removing people from the list is a higher-leverage activity than ranking them.
A checklist, and where it breaks
The working checklist: median and spread of price, shape of the price histogram, total SKU count, items per top category, product-to-variant ratio, title and description consistency, and keyword count divided by product count. Seven reads, all mechanical, all doable in bulk. Run them across a filtered segment and you have a prioritised list with a specific opening line attached to each row.
Where it breaks: productCount is a snapshot and some platforms surface variants as separate products, which inflates the number. Dropshipping catalogs can be enormous and mean nothing. Business-to-business stores frequently hide prices behind a login, which makes the price read impossible rather than merely noisy. And a catalog can be huge because the merchant imported a supplier feed once and forgot about it.
So treat the catalog read as a hypothesis you test in the first two lines of the email, not a verdict you deliver. The merchant will correct you if you are wrong, and a correction is a reply, which is the entire point of the exercise.