chivvy. Start a conversation
What we doWorkHow we workGuidesArticlesWho we areFree toolContactStart a conversation
What we do

Knowing which customers
are slipping away

Almost nobody fires you. They order less, then later, then not at all. The signal is already in your order history — we build the thing that watches for it.

Start a conversation See our work
The problem

Ask why a good customer was lost and you'll usually hear about price or a competitor. Ask when anyone noticed and the answer is almost always: when the orders had already stopped.

Customer loss is a slope, not an event — smaller orders, longer gaps, a narrower range — and it's recorded in perfect detail in your own systems for months beforehand. Nobody spots it because nobody can hold the ordering rhythm of two hundred accounts in their head.

What we build
01

Cadence against their own history

Each customer measured against their own median gap between orders, not an all-customer rule that screams about quarterly buyers and misses weekly ones.

02

Volume trend

The accounts still ordering on time but at half the size — invisible to any 'days since last order' report.

03

Range breadth

How much of you they buy. An account down from four lines to one is fragile even when the revenue looks stable.

04

A weekly call list

Who to ring, why now, what they usually buy, and what similar customers take that they don't — so the call has a purpose beyond checking in.

05

Seasonality handled

Compared like-for-like against the same weeks last year, and against the drift of your whole book, so a quiet November doesn't look like a crisis.

06

A short list, deliberately

Capped at the handful that matter this week. A list of forty gets ignored exactly as fast as a forty-tile dashboard.

Proof

Staleness detection in practice

Every stage in the pipeline has a threshold, and anything past it surfaces as 'needs a nudge' with the reason attached — the difference between recording what happened and prompting someone to act.

altropartners.uk
Staleness detection in practice
Case study

Software for a firm and its partners

Referrals between a financial services firm and its partner accountants, tracked end to end. Each firm logs into what looks like their own software.

Read the case study →
Straight answers

The things people ask.

What data does this need?

Order or invoice history with customer, date and value — which every business has. More detail makes it better, but that's the minimum.

How far back do you need?

Ideally two years, so seasonality is visible. One year works; less makes the comparisons weaker.

Does it work for service businesses?

Yes — the pattern is the same whether it's orders, bookings, jobs or repeat engagements. What changes is the rhythm being measured.

Will it just create a long list of false alarms?

That's the usual failure, and it comes from comparing everyone against one rule. Comparing each account against its own history, excluding anyone with too few orders to have a pattern, and capping the list is what keeps it credible.

Can it contact them automatically?

It can draft the approach. Sending stays with a person — these are your relationships, and an automated message to a good customer who's quiet for a legitimate reason does real damage.

Read more

Written on this, at length.

Guide

Your best customers go quiet long before they leave

Read →
Guide

Getting paid faster without becoming the bad guy

Read →
Article

Small businesses don't have a data problem. They have a joining-up problem

Read →
Get started

Tell us what's eating your week.

Tell us the job that keeps not getting done. We'll tell you if software can take it off you, and what it costs.

Start a conversation