--- title: "The long tail of small accounts: the hidden cost" description: "Small accounts bring little revenue and almost all the admin work. Why that happens, why handing them off costs more than it looks, and what can be automated." url: https://studiochat.io/blog/la-cola-larga-de-clientes locale: en translations: { es: https://studiochat.io/es/blog/la-cola-larga-de-clientes.md, pt: https://studiochat.io/pt/blog/la-cola-larga-de-clientes.md } site: https://studiochat.io --- [ Back to blog](/blog) There is a conversation that keeps coming up with companies that sell products to other companies, and it almost always starts the same way: **it is not that we are short of customers, it is that the small ones eat the whole week**. The number behind it is usually the same too. Twenty percent of the customers account for most of the revenue. The remaining eighty percent contribute a small slice of the volume and take up most of the day for the admin team and the sales reps. The interesting part is not the number, which is old and well known. The interesting part is that almost nobody has it measured in hours. ## [Why the work does not scale with the amount](#why-the-work-does-not-scale-with-the-amount) Revenue scales with order size. Admin work does not. Taking a one thousand dollar order and taking a one hundred thousand dollar order are, in minutes, nearly the same task. Someone reads the message, works out which customer it belongs to, confirms the current price, checks stock, builds the order, sends it for confirmation, invoices it and then chases the payment. Every one of those steps is paid per order, not per amount. So the arithmetic flips. If eighty percent of the customers place eighty percent of the orders, then eighty percent of the admin hours are serving the part of the book that bills the least. The company does not have a sales problem: it has an operational arithmetic problem that was never written down anywhere. It is worth looking at this with your own numbers, because the average in an article describes nobody's actual customer book. These two bars are the same book seen two ways: on top where the revenue comes from, underneath where the hours come from. The long tail, with your numbers Active customers: 180 Volume from the largest 20%: 80% Orders a month from a large customer: 6 Orders a month from a small customer: 2 Admin minutes per order: 12 The tail is 144 customers: they give you 20% of revenue and take 57% of the hours. For every point of revenue, a tail customer costs you 5.3 times more hours than a large one. Not because they are harder to work with: because the admin work per order is almost the same whatever the order is worth. Which is why the obvious way out, handing the tail to a distributor, is paid for with something that was not in the calculation: you stop deciding what gets sold to 144 customers. Every number here is yours. No StudioChat customer data is used. The gap between the two bars is the problem. If it is small, the tail was not the issue and it is worth looking elsewhere. If it is large, it is already in hours and in money, and it can be taken into a meeting. The numbers are whatever you put in: there is no measurement of a client of ours here. This object also has [a page of its own](/your-long-tail-of-customers), in case you want to send just that. ## [The obvious way out has a price that is not in the calculation](#the-obvious-way-out-has-a-price-that-is-not-in-the-calculation) The first answer that comes up is always the same: hand the tail to a distributor. It works for the operational problem. It takes away the small orders, the small collections and the small follow ups. But it takes two things that were never on the spreadsheet. The first is what gets sold. A distributor sells what suits their own margin, not what the company wants to push this quarter. If there is a new product, a premium line or seasonal surplus, the distributor has no reason to prioritise it. The second is who it gets sold to. The company stops knowing who those customers are, what they buy and when they stop buying. The day one of them grows, it is no longer a customer of yours. So the right question is not whether to let the tail go, but how many hours letting it go actually gives back. That calculation can be done. ## [What can be removed without letting the customer go](#what-can-be-removed-without-letting-the-customer-go) Between handling every order by hand and handing the book to a third party there is a good deal of room. A recurring order from a known customer is, almost always, a conversation with a fixed shape. The customer writes on the channel they already use, asks for what they usually ask for, at prices already agreed. That conversation can be taken without anyone reading it line by line: identify the customer by their number, build the draft order against the current price list and leave it ready to confirm. What stays on the other side is the decision: the off policy discount, the exception, the credit limit, the customer asking for something they have never bought. There is still a person there, and it is a good idea that there still is. The difference between the two halves is not technical. It is that one half is work and the other half is judgement. ## [What this decides](#what-this-decides) The argument about the tail rarely stalls for lack of reasoning. It stalls because nobody put the hours next to the revenue, and without that number the conversation stays at the small ones are a lot of work. With the number, the decision becomes concrete: how much of that load can be taken away without letting the customer go, and what letting them go would really cost. Frequently asked questions ### Why does a small customer cost almost the same as a large one? Because admin work is paid per order, not per amount. Taking the order, confirming the price, checking stock, issuing it and chasing payment take nearly the same minutes whatever the order is worth. A customer who orders little and often creates almost the same load as one who orders a lot and often. ### Should the small accounts go to a distributor? It removes the operational load, but it takes two things with it: control over which products those customers are offered, and the direct relationship. Many companies find out late that a distributor sells what suits the distributor, not what suits the brand. Before deciding, it is worth calculating how many hours actually come back. ### Which part of order taking can an AI agent handle? The repetitive, verifiable part: receiving the order on the channel it already arrives through, identifying the customer, drafting the order with current prices and leaving it ready for someone to confirm. What should not be automated without deciding it first is the discount, the exception and the credit limit. Share [ X](https://x.com/intent/post?text=80%25%20of%20your%20customers%20give%20you%2020%25%20of%20the%20volume%2C%20and%20eat%20the%20whole%20week&url=https%3A%2F%2Fstudiochat.io%2Fblog%2Fla-cola-larga-de-clientes) [ LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fstudiochat.io%2Fblog%2Fla-cola-larga-de-clientes) [ WhatsApp](https://wa.me/?text=80%25%20of%20your%20customers%20give%20you%2020%25%20of%20the%20volume%2C%20and%20eat%20the%20whole%20week%20https%3A%2F%2Fstudiochat.io%2Fblog%2Fla-cola-larga-de-clientes)Copy link Ignacio Puig Moreno Ignacio Puig Moreno · Co-founder · Business, StudioChat Want agents like these working for your team? [Talk to us now](https://cal.com/studiochat)