Accounts Don't Churn. Lines Do — and Yours Are Leaking Right Now

Distribution Strategy Group keeps re-asking a version of the same question — how many customers did you lose, and which ones were worth keeping. Wrong altitude. Accounts almost never churn in wholesale distribution; they just buy less of you, one category at a time, until the number on the P&L finally admits what the buyer decided two years earlier. If you're only measuring at the account level, you are measuring the funeral, not the illness.

The account is the last thing to move

Here's the mechanics nobody puts in the churn deck. A regional MRO buyer doesn't wake up one Tuesday and switch distributors. They start with one category — say, fasteners, because your site can't confirm a thread pitch or a coating spec fast enough and a competitor's PDP shows it in one screen. They test the competitor once. It works. Six months later, safety supplies move the same way, because that category has the same findability problem on your site. A year after that, the buyer is single-sourcing electrical from you and everything else from two other vendors and Amazon Business. Your CRM still shows an active account, on-time payments, a rep who "has a great relationship" there. Revenue per account has been sliding for eighteen months and everyone attributes it to "the market."

This is the account that eventually does churn — but by the time it does, you've already lost the categories that made it profitable. The account-level number that finally shows up in a churn report is a lagging indicator of a decision the buyer made SKU by SKU, quietly, without ever telling your sales rep.

Where the trade press has been looking

Distribution Strategy Group has argued that the churn number worth tracking isn't total customer count but profitable customer count — that most distributors can't tell a valuable account from a mediocre one because their ERP reports monthly aggregates instead of order-level margin, and that "your most valuable customers can look just like everyone else until it's too late." That's a real improvement over the 2023 piece asking distributors to simply count losses, and it echoes a 2021 argument that most churn programs fail because they react to a departure instead of predicting one.

We agree with the diagnosis and think it still stops one layer too shallow. Order-level profitability is a better lens than account-level revenue, but it's still an account-level metric rolled up from orders — it tells you an account got less valuable, not why, and not which piece of its business already walked out the door. A distributor can build a flawless A/B/C/D profitability model, watch a mid-tier account's order value erode for a year, and still have no idea that the erosion is 100% concentrated in three product lines the buyer now sources from a competitor with a better spec table. The fix DSG proposes — tighter risk buckets, faster escalation — treats the symptom at the account. The disease lives at the category.

The metric that's missing: category share per account, over time

This isn't exotic. Most distributors already have the data to build it; they just haven't pointed a report at it. For every active account, track share of that account's spend by product category or line, quarter over quarter, against the account's own trailing baseline — not against total company revenue, which hides the signal in noise. A account that's flat in total dollars but has quietly dropped from 40% to 12% share in one category isn't a stable account. It's a defecting one, and it's telling you exactly where.

This is a materially different report than an at-risk dashboard built on order frequency or days-since-last-purchase, which is what most CRM-adjacent churn tools default to. Frequency and recency catch accounts that are already gone. Category share catches accounts that are still buying from you everywhere except the one place a competitor got there first — which is the only stage where a sales call, a price adjustment, or a catalog fix can still reverse it.

Why the leak so often starts on the website, not in the field

The behavioral research backs up where these defections start. In Sana Commerce's 2025 B2B buyer survey, 85% of B2B buyers reported frustrations that led them to abandon a purchase, and 75% said those frustrations made them consider switching suppliers entirely — not because of price or product, but because of the buying experience itself. Separate research on switching behavior puts the number even higher: 44% of B2B buyers say they're willing to switch suppliers purely because the digital buying experience doesn't meet expectations, independent of the underlying product or price. And Amazon Business's discovery layer — now north of $60 billion in annualized sales — is explicit that suppliers who can't surface complete, structured product data lose the buyer's attention "regardless of catalog depth or account relationship history."

Put those together and the mechanism is obvious. A buyer with a relationship, a payment terms agreement, and a rep on speed dial will still quietly route a category away from you if your site can't confirm the spec, show the fitment, or even surface the SKU in search — because the alternative is one tab away and doesn't make them call anyone. The relationship doesn't die. It just stops covering that category.

Measuring the leak is only half the job

Line-level share tracking tells you which accounts are defecting and where. It doesn't tell you why your site loses that category to a competitor's in the first place — and the answer is almost always product data, not price. Search that can't handle a part number variant. A spec table with three of the eleven attributes a buyer needs to confirm fit. No structured data for the crawlers now doing agentic product discovery on a buyer's behalf. This is exactly the surface we built the Digital Readiness Index to measure across 200+ distributors in the Top Distributors 2026 index — search depth, spec completeness, structured data, and the other signals that determine whether a category is easy or hard to buy from you.

You don't need to rebuild your PIM to close that gap. Anglera sits on top of whatever system already holds your product data and does the enrichment work — filling spec gaps, fixing fitment data, making the catalog machine-readable — typically live within a few weeks. Silent churn is a product-data problem for a long time before it becomes a relationship problem. Measure it there first.