Notes for owners · Channel and customer growth
What is customer health scoring?
A health score is a small set of signals, as few as five on a sheet, that shows which customers are drifting before they tell you.
The GullySales team · Updated 15 Sept 2026 · 7 min read
On this page
Customer health scoring means picking a handful of signals that reliably show whether an existing customer is doing well with what they bought. You record those signals for every account and add them up into one number. The point is to find the accounts that are drifting before the renewal conversation, the complaint or the silence that turns into a cancellation. For most Indian SMEs the whole thing is a spreadsheet with five columns, reviewed once a month, and it is worth more than any retention campaign.
Why a score beats a feeling
Every owner believes they know which customers are happy. The belief is usually wrong in a predictable direction: the loudest customers are assumed to be the unhappy ones, and the quiet ones are assumed to be content.
Silence is the more dangerous signal. A customer who complains is engaged and reachable. A customer who has stopped calling, stopped ordering spares and stopped attending the review meeting has already begun leaving. Nobody has noticed, because there is no complaint on any list.
A score forces the quiet accounts onto the same page as the noisy ones. That is its entire value.
The five signals most SMEs should use
Choose signals you can actually observe every month without a project.
Usage or consumption. Are they using what they bought? For a machine, the service call and spares pattern. For a software or service retainer, logins or work requested. For a supply relationship, order frequency against their normal pattern.
Recency of a real conversation. Days since somebody from your side spoke to them about something other than a problem or an invoice. This is the single best predictor we see in small businesses.
Payment behaviour. Whether the delay is lengthening matters more than whether they eventually pay. A customer moving from 30 days to 60 is making a decision about you, and has not said so.
Relationship depth. How many people at their end know you, and is the person who chose you still there. A single-contact account where the champion has resigned is at high risk regardless of how well the work is going.
Open issues and their age. Not the count, the age. Three issues closed in a day is a healthy account. One issue open for six weeks is a churn event in progress.
A scoring sheet you can build this week
| Signal | Green (2 points) | Amber (1 point) | Red (0 points) |
|---|---|---|---|
| Usage or order pattern | At or above normal | Down noticeably | Stopped |
| Days since a real conversation | Under 60 | 60 to 120 | Over 120 |
| Payment behaviour | Within terms | Stretching | Overdue and disputed |
| Relationship depth | Three or more contacts, champion in place | Two contacts | One contact, or champion gone |
| Oldest open issue | None, or under a week | Two to four weeks | Over a month |
Ten points is healthy. Six or seven means the account owner books a conversation within a fortnight. Five or below and the owner of the business makes the call personally.
Weight the signals if you must, but not at the start. Equal weights and honest data beat clever weights and guessed data every time.
A Bengaluru facility management firm, and the contract that was already gone
For example, imagine a facility management company operating housekeeping and security contracts for offices and apartment complexes around Whitefield and Electronic City. The details and figures below are illustrative. Revenue looked stable, and two contracts were lost in the same quarter, both a surprise to the management.
Scoring the portfolio afterwards showed both had been red for months on the same two signals. The facility manager who had originally appointed them had changed at both sites, so relationship depth was one contact and the champion was gone. And neither had had a review meeting in over six months, because the operations were running smoothly and nobody saw a reason to visit.
The sheet was then built for all the live contracts. Illustrative figures: on a portfolio of thirty-four contracts averaging ₹1.4 lakh a month, eleven scored six or below, and nine of those eleven had no scheduled review meeting. Two supervisors were given eight accounts each to visit monthly, with a written note of what the client said. The score was not a prediction engine. It was a way of deciding whose Tuesday morning goes where.
How to run it without it becoming a reporting exercise
Update it monthly, on the same day, and give the job to one person. Ten minutes per account at most.
Review only the accounts that moved. A green account that stayed green needs no discussion, and a meeting that walks through the whole list will be abandoned by the third month.
Attach an action to every red and amber, with an owner and a date, and check it next month. A score with no consequence is a colour chart.
And read the pattern once a quarter. If a particular signal keeps turning red before a customer leaves, weight it higher. If a signal never predicts anything, replace it. The score should be yours within two quarters, not the version you copied from an article.
Where this connects to new business
A portfolio with visible health has a second benefit beyond retention. The green accounts are where case studies, testimonials, reviews and referrals come from, and those are the only proof new buyers actually believe. Most SMEs ask for a testimonial from whoever is handy rather than from whoever is genuinely doing well, and the difference shows in what they get back.
What to do next
List every customer with repeat revenue and fill in just two columns: days since a real conversation, and whether the person who chose you still works there. That alone will identify the accounts at risk in most small businesses, and it takes an afternoon. Add the other three signals next month.
