You can name every customer you won. Now name the ones you quietly lost.
Gully Sales measures your churn by cohort, finds the reasons customers actually leave, and builds the triggers and save plays your team runs while an account can still be kept.
- A churn number you can defend, measured by cohort rather than by feeling.
- The real reasons customers leave, taken from their words and your own records.
- Save plays your team runs on a trigger, before the account is already gone.
Gully Sales Private Limited works with businesses across India, and every finding on your page comes from your own customer records.
In one paragraph
What is Churn Analysis and Reduction?
Churn analysis and reduction is the work of measuring how many customers you lose, understanding why, and changing what causes it. Gully Sales builds a cohort baseline from your own records, ranks the drivers behind each loss, sets the triggers that flag an account early, and writes the save plays your team runs in response.
The problem
You lost customers last year, and nobody can say how many or why.
Churn is the one number in a business that nobody is responsible for. Sales is measured on what it brings in. Delivery is measured on what it ships. Nobody is measured on what quietly left. So the customer who stopped ordering in March is noticed in September, the reason recorded is price because price is what they said, and the loss goes into the same folder as the last one. None of this is negligence. It is what happens when a number was never defined, so it was never owned.
You will recognise it as
- You can quote last year's sales figure to the rupee, but not how many customers you started the year with and still had at the end.
- A customer stops ordering and it is noticed months later, when somebody happens to open their account.
- The reason recorded for every departure is price, because price is what they said and nobody asked a second question.
- Two people in your team would give two different answers to how bad churn is, and both would be guessing.
- Revenue is flat while your sales team adds new customers every month, and nobody has joined those two facts together.
- The attempt to save an account happens after the customer has already decided, usually as a discount.
What it costs the business
- New sales are spent replacing revenue you already had, so the business works harder every year to stand still.
- Acquisition cost keeps climbing, because each new customer has a shorter life than the one they replaced.
- Discounting becomes the only retention tool you have, and margin falls fastest on the accounts you most wanted to keep.
- Product, service and delivery decisions get made without the one piece of evidence that says what drove customers away.
- Planning becomes unreliable, since the base you forecast from leaks at a rate nobody has ever measured.
Why it persists. It persists because churn is easy to avoid measuring. Every business has a revenue figure on a screen; almost none has a retention curve. Without a definition, a distributor cannot even say when a customer has left, so there is nothing to count. It persists because the reason a customer gives at the end is polite and short, and taking it at face value is easier than asking a second question. And because losses arrive one at a time, each one feels like a special case rather than a pattern.
If it stays unresolved. Left alone, the leak sets the pace of the business. Every year your sales team has to win the same customers again before it wins any new ones, acquisition costs rise, and the accounts you do keep are increasingly the ones you kept with a discount. The day somebody asks what a customer is worth over their whole life with you, nobody will be able to answer.
What changes
Churn becomes a number you manage, not a story told afterwards.
In the first weeks
- A stated definition of churn for your business, so everyone counts the same thing the same way.
- A cohort baseline showing how many customers you kept, lost and won back across the periods your data covers.
- The reasons behind recent losses, written down from your records and from customers who agreed to talk.
In how the work runs
- Named triggers that flag an at-risk account, each with a threshold, an owner and a response time.
- Save plays your team can run without asking permission, matched to the reason the account is at risk.
- A renewal and check-in rhythm, so the first conversation about continuing is not also the last one.
- Exit questions and a loss-reason field that make every future departure informative.
In sales and marketing
- Fewer accounts lost for reasons you could have handled, because the signal reached somebody in time.
- Less margin given away, because a discount stops being the only save your team knows.
- A reactivation offer aimed at lapsed customers whose objection has since been fixed.
- Acquisition spending judged against repeat revenue, not against first orders alone.
In what management can see
- A churn report leadership actually reads: by cohort, by segment and by reason, refreshed each month.
- A clear view of which segments hold customers and which quietly lose them.
Over the longer term
- Lifetime value read by segment, so you can see which customers are worth acquiring more of.
- A retention habit that survives staff changes, because the definitions and plays are written down.
We control the analysis, the definitions, the triggers, the plays and the reporting rhythm, and we run them with your team. Whether any one customer stays also depends on your product, your service and your pricing. Where the cause of churn sits outside what this work can change, we will say so.
Who it is for
This is for businesses that have lost customers without ever counting them.
The businesses it suits
- Service businesses on retainers, contracts or annual maintenance, where one non-renewal costs a year of revenue.
- Subscription and software businesses whose growth is being eaten by cancellations they cannot explain.
- Distributors, dealers and B2B suppliers with repeat-order customers who stop ordering without ever saying so.
- Clinics, gyms, institutes and membership businesses where attendance falls long before anybody formally leaves.
- Companies with a small customer base, where losing four or five accounts changes the whole year.
- Businesses with history sitting in a CRM, billing tool or order book that has never been read for churn.
What usually prompts the call
- Revenue is flat although your sales team is adding new customers every month.
- A large account left recently and the post-mortem produced opinions rather than evidence.
- Renewal season arrives as a surprise, and the negotiation always finishes in a discount.
- You are planning a price revision and do not know how much of the base would stay.
- An investor, lender or board has asked for your churn rate and you could not answer confidently.
- Support and delivery keep hearing the same complaint, and nobody has connected it to the customers you lost.
What Gully Sales does
The work, component by component.
Churn definition and cohort baseline
We agree what churn means in your business, then count it. Customers are grouped into cohorts by when they joined, and each cohort is followed period by period: how many stayed, how many left, and how much revenue went with them. Revenue retention is kept separate from customer counts, because a business can lose customers and still grow revenue, or keep customers and still shrink.
- Why it matters:
- Without a shared definition and a baseline, every later claim of improvement is unprovable and every argument about churn is an argument about opinions.
- You receive:
- A cohort retention table and a one-page churn baseline, with the counting definition written beside it.
- Business value:
- You get a retention number you can put in front of a board, a lender or your own team and defend.
Loss reason analysis and risk drivers
Every loss we can reconstruct is examined: what the account looked like before it went, what changed, what was said and what your team saw at the time. Where you allow it, we speak to former and current customers directly. The reasons are then grouped into drivers and ranked by how much revenue each one costs you, so the list stops being a collection of anecdotes.
- Why it matters:
- The reason a customer states at the end is rarely the reason the relationship failed, and acting on the stated reason wastes the effort.
- You receive:
- A ranked driver list with the evidence behind each one, plus an anonymised extract of interview notes.
- Business value:
- Your fixes get aimed at what actually costs you customers, in the order that costs you most.
Early warning triggers and thresholds
For each driver we find an observable signal that appeared before the loss: an order gap, a usage drop, an unresolved ticket, a missed visit, a first late payment, a changed contact person. Each becomes a trigger with a threshold, an owner and a response time. Where your tools allow it the trigger fires automatically; where they do not, it becomes a weekly list somebody reviews.
- Why it matters:
- A risk nobody sees is not managed, and a risk seen by nobody in particular is not managed either.
- You receive:
- A trigger sheet naming the signal, the threshold, the owner, the response time and where the flag appears.
- Business value:
- At-risk accounts reach a named person while there is still time to do something about them.
Save plays and intervention playbook
For each trigger there is a written play: who contacts the customer, what they say, what may be offered, what has to be fixed inside your business first, and what happens when the customer does not respond. Plays are matched to the driver, so a service failure gets a repair and a review meeting, while a departed champion gets a fresh relationship built with the person who replaced them.
- Why it matters:
- Teams hesitate when a customer goes quiet, because nobody has told them what they are allowed to do about it.
- You receive:
- A save playbook with call and message wording, an escalation path and an approval limit for each play.
- Business value:
- Your team acts on the day the flag appears instead of waiting for somebody to grant permission.
Renewal and check-in cadence
We set the rhythm that keeps the relationship in view between purchases: when the first renewal conversation happens, what is reviewed in it, which accounts get a scheduled check-in and what evidence of value is carried into the meeting. The cadence is sized to the team you actually have, so it survives a busy month rather than lapsing in one.
- Why it matters:
- A renewal argued in its final week is argued on price. A renewal prepared across the term is argued on what was delivered.
- You receive:
- A renewal calendar and a check-in agenda template, with the account evidence to assemble beforehand.
- Business value:
- Renewal decisions get made early and calmly, and the surprise cancellation becomes rare.
Reactivation offer for lapsed customers
Customers already lost are sorted by why they left and whether that reason still holds. Those whose objection has since been addressed get a specific, honest approach that says what changed. The rest are left alone. We write the segments, the offer and the sequence, and set a rule for when a lapsed customer is retired from the list for good.
- Why it matters:
- A customer who left over something you have since fixed is the warmest name in your database, and the easiest to insult with a generic mailer.
- You receive:
- A segmented lapsed-customer list, the offer written for each segment, and the outreach sequence.
- Business value:
- Revenue you had written off gets a fair second approach, and everybody else stops being mailed for no reason.
Impact tracking and churn reporting
One report, produced the same way every month: churn and retention by cohort and segment, accounts flagged, plays run, saves and losses, revenue retained, and the reasons recorded during the month. It is built so somebody in your team can produce it in an hour without us, and so this month can be compared honestly with the baseline.
- Why it matters:
- A retention programme that is not reported reverts to instinct within two quarters, whatever was agreed at the start.
- You receive:
- A monthly churn report template, the calculation notes, and one month filled in as a worked sample.
- Business value:
- Leadership sees whether retention is improving, and the team sees which plays are worth running.
What you will have at the end.
- A written churn definition agreed with you, and the counting rules that go with it
- A cohort retention table built from your own customer, order and billing records
- A one-page churn and retention baseline you can circulate to your leadership team
- A ranked list of churn drivers, each with the evidence that put it in that position
- Anonymised notes from conversations with current and former customers, in their words
- A trigger sheet: signal, threshold, owner and response time for each risk worth watching
- A save playbook with call and message wording, escalation path and approval limits
- A renewal and check-in calendar sized to the team you actually have
- A segmented lapsed-customer list with a reactivation offer written for each segment
- A monthly churn report template, with one month completed as a worked sample
- A short working session so your team can run the plays and the report without us
- A prioritised action list separating the quick fixes from the structural ones
How it runs
The engagement, step by step.
- 1
Records pull and churn definition
We start with what you already have: customer list, start and end dates, orders or invoices, CRM notes, support tickets. Then we agree what a lost customer means in your business, because a distributor, a clinic and a subscription business each lose customers in a different way. The definition is written down before a single number is produced.
- You provide:
- Access to your customer, billing and support records, and an hour with whoever knows them.
- We produce:
- A written churn definition, an inventory of the data available, and a note of what is missing.
- Done when:
- You and we count churn the same way, using the same words.
- 2
Cohort baseline and segmentation
Customers are grouped by when they joined and followed forward, so retention reads as a curve rather than one annual figure. We then break the base by segment — size, product, region, channel, acquisition source — and look for where the curve falls away fastest. This is where most businesses first see that their churn is concentrated, not general.
- You provide:
- Confirmation of how you group customers commercially, and any segments you already work with.
- We produce:
- A cohort retention table, segment breakdowns and the baseline one-pager.
- Done when:
- You can say how many customers and how much revenue you retained, cohort by cohort.
- 3
Loss reason research
We reconstruct recent losses from the records and from your team, then test what we find with customers. Former customers are asked why they left. Current ones are asked what would make them consider leaving. Answers are recorded in their own words, not summarised into ours. Records show what happened; conversations explain it, and the two rarely agree at first.
- You provide:
- Introductions to a small number of current and former customers, and time with sales and service staff.
- We produce:
- Loss case reconstructions and anonymised interview notes, quoted rather than paraphrased.
- Done when:
- The reasons behind your losses exist as written evidence rather than as opinion.
- 4
Driver ranking and trigger design
Reasons are grouped into drivers and ranked by the revenue each one costs. For every driver we look back through the data for a signal that appeared before the loss, then set a threshold that would have caught it without flooding your team with false alarms. Each trigger is given an owner and a response time, so the flag lands somewhere specific.
- You provide:
- A decision on who owns each trigger, and an honest view of what your tools can automate today.
- We produce:
- The ranked driver list and the trigger sheet with thresholds, owners and response times.
- Done when:
- Every driver worth acting on has a signal, a threshold and a name against it.
- 5
Save plays, cadence and reactivation offer
We write the response for each trigger: the contact, the message, what may be offered, the internal fix and the escalation. The renewal and check-in rhythm is set around them. Lapsed customers are sorted by reason, and an offer is written only for the segments whose objection has genuinely been addressed since they left.
- You provide:
- Your approval limits, what can be offered commercially, and who is allowed to offer it.
- We produce:
- The save playbook, the renewal and check-in calendar, and the reactivation segments and offer.
- Done when:
- Your team knows what to do on the day a flag appears, without escalating first.
- 6
Run the first cycle with your team
The programme goes live and we run it beside you. Flags are reviewed together, plays are used on real accounts, and the wording is corrected from what customers actually say back. Thresholds that fire too often are tightened; those that fire too late are loosened. This is the stage where a document becomes a habit, and it is the stage most retention work skips.
- You provide:
- Your team's time in a short weekly review, and honesty about what is not being done.
- We produce:
- Corrected thresholds and plays, a record of every flag and its outcome, and coaching for the team.
- Done when:
- Flags are being actioned by your own people within the agreed response time.
- 7
Measure, review and hand over
We compare the live period against the baseline, read what the saves and the losses are telling you, and update the driver list with every reason recorded since the start. Then the programme is handed over properly: the report is produced by your team, the playbook has a named owner, and the review sits in somebody's calendar.
- You provide:
- A named owner for the programme and a decision on the review rhythm you intend to keep.
- We produce:
- The first impact read against baseline, an updated driver list, and the handover pack.
- Done when:
- Your team runs the programme and produces the monthly report without us.
Ways to work with us
How this work usually begins.
Churn diagnostic
A focused analysis of your records: a churn definition, a cohort baseline and a first ranked view of why customers are leaving. Suitable when you need the number and the reasons before deciding what to build around them.
Churn reduction programme build
The full engagement: baseline, customer research, ranked drivers, triggers and thresholds, save playbook, renewal cadence, reactivation offer and the reporting template, built with the people who will run it.
Build plus first-cycle support
The programme build followed by hands-on support through the first cycle. Flags are reviewed together, plays are run alongside your team, and thresholds are corrected from what actually happens on real accounts.
Quarterly churn review
A standing review once the programme is live. New losses are examined, drivers are re-ranked, plays are judged on what they held, and the report is read with your leadership team rather than filed.
Why Gully Sales
What you are actually choosing when you choose us.
We count before we advise.
Every recommendation starts from your own records. Where the data cannot support a claim, we say so rather than filling the gap with an industry average borrowed from a different market.
Sales and service are read as one system.
Most churn is created long before a complaint arrives, in what was promised at the sale or missed at delivery. We work across both, because studying support alone finds the symptom and leaves the cause in place.
We separate what you can change from what you cannot.
Some customers leave because they shut down, moved city or were acquired. Naming those losses honestly keeps your effort and your budget on the accounts you could genuinely have kept.
Written for the people who will run it.
The playbook is short, plain and specific to your business, so a salesperson in their second week can pick it up. Nothing in it depends on a consultant staying in the room.
We say when the answer is smaller than the project.
If your churn traces back to one broken process, we will tell you to fix that and stop, rather than sell you a programme you do not need yet.
Where it applies
The same service, in different businesses.
Industrial distribution and B2B supply
- The situation:
- A supplier with a few hundred trade accounts watched revenue stay flat while the sales team added customers every month. Nobody was tracking the accounts that had quietly stopped ordering.
- How it applies:
- Churn was defined as an order gap longer than the account's own normal cycle. The cohort baseline showed where the loss sat, and an order-gap trigger put lapsing accounts on a weekly list with an owner against each.
- Likely benefit:
- Lapsing accounts get a call while the relationship is still warm, and the flat revenue finally has an explanation the owner can act on.
Clinics and healthcare
- The situation:
- A multi-doctor clinic filled its appointment book but saw the same patients less and less often. Follow-ups were booked verbally at the desk and never chased when they were missed.
- How it applies:
- The missed follow-up became the primary signal, and the recall call was written as a play with an owner, a script and a time limit. Reasons for dropping treatment were captured at the desk in the patient's own words.
- Likely benefit:
- More patients complete the course of treatment, and the clinic can see which conditions lose people midway and why.
Software and subscription businesses
- The situation:
- A small software company knew its monthly cancellation figure but not which customers were about to join it. The cancellation form said price, while the usage data told a different story.
- How it applies:
- Cohorts by signup month showed churn concentrated in one acquisition channel. Login and feature-use drops became triggers, and a save play was written for each driver instead of one blanket discount offer.
- Likely benefit:
- Cancellations get addressed at their cause, and the weakest acquisition channel is repriced or paused on evidence rather than on instinct.
Education, coaching and training
- The situation:
- An institute enrolled well but lost a share of every batch before completion, and each student who left also took with them the referrals they would have made.
- How it applies:
- Attendance gaps and missed assignments became the early signal. A structured call at the first missed session replaced the end-of-term post-mortem, and exit reasons were recorded batch by batch.
- Likely benefit:
- More students finish, and the reasons for dropping out are known while the batch is still running and can still be changed.
Facility, maintenance and annual contract services
- The situation:
- An annual maintenance provider renewed most contracts but negotiated every one on price in the final week, with no assembled record of what the past year had actually delivered.
- How it applies:
- A renewal calendar started the conversation months earlier, and each account's service record was assembled as evidence. Contracts at risk were flagged from complaint and visit data well ahead of the due date.
- Likely benefit:
- Renewal conversations open with what was delivered, and fewer contracts have to be rescued with a discount at the last minute.
Consumer brands with repeat purchase
- The situation:
- A direct-to-consumer brand acquired buyers steadily but second orders were rare, and the team could not say whether the problem was the product, the delivery or the follow-up.
- How it applies:
- Repeat-purchase cohorts by month showed exactly where buyers dropped off. Post-delivery feedback and a first-repeat window trigger separated product complaints from simple absence of follow-up.
- Likely benefit:
- Acquisition spending is judged on repeat revenue rather than first orders, and the real reason for one-time buying is finally named.
Questions buyers ask
Before you enquire, the answers you will want.
How early can we see that a customer is at risk?
That depends on what your records already capture. In most businesses the signal exists well before the cancellation: a drop in order frequency, a missed service visit, a support complaint that was never closed, a champion who left the company, an invoice paid late for the first time. We find which of these actually preceded your past losses, set a threshold for each, and give the flag an owner and a response time.
How long does this engagement take?
It depends on how many customers you have, how clean your records are, and whether we are analysing churn only or also building the save plays and running the first cycle with your team. A diagnostic on a small customer base moves quickly. A full programme with customer interviews takes longer. We scope it after seeing your data and agree the sequence before starting, rather than quoting a period we cannot hold to.
What do we need to give you to start?
A customer list with start dates and, where relevant, end dates. Order or invoice history. Whatever your CRM, billing tool or order book already holds. Access to the people who deal with customers daily, in sales, service and accounts. Permission to speak with a small number of current and former customers. If some of this does not exist, we say so and build the baseline forward from the month we begin.
How is success measured?
Against the baseline we record before anything changes. The main measures are churn rate by cohort and segment, renewal rate, revenue retained, accounts flagged and saved, and lifetime value once enough cycles have passed to read it fairly. We also track adoption, satisfaction and support resolution, because those move first. Every number is defined in writing, so it means the same thing next quarter as it did in the first month.
What is not included in this work?
We do not run your support desk, negotiate your renewals for you or rebuild your product. We do not buy or install a new CRM as part of this; we work with what you have and say plainly when a tool is genuinely holding you back. Helpdesk design, contact centre setup and loyalty scheme mechanics are separate pieces of work, and we will point you to them rather than fold them in quietly.
We have only forty customers. Is churn analysis worth it?
Often it matters more. With a small base, one departure is a measurable share of your revenue, and the pattern behind your losses is usually visible without statistics. The work simply looks different: fewer cohorts, more conversations, and judgement applied account by account rather than by segment. What you get is the same, which is a defensible number, named reasons and a response your team runs before the account has gone.
How is this different from customer health scoring?
Health scoring gives you one score per account so you can see who to worry about today. Churn analysis looks backwards across the whole customer base to work out what actually caused past losses, then forwards to reduce those causes. The two fit together, because the drivers this work identifies are exactly what a health score should be built from. Many businesses start here and add scoring afterwards.
Will you speak to customers who have already left us?
With your permission and your introduction, yes, and it is usually the most useful hour of the whole project. Former customers answer honestly because they have nothing left to negotiate. We ask a short, neutral set of questions, record the answers in their own words and share them unedited. If you would rather we did not approach them, we work from records and from your team's account of each loss instead.
4 more questions
Can churn be reduced without cutting prices?
In most cases the discount is treating the wrong problem. When a customer leaves over a service failure, a stalled implementation or a contact who changed jobs, price was the reason they gave, not the reason they went. Save plays are built around the real driver: a repair, a review meeting, a training session, a different point of contact, a changed delivery schedule. Discount stays available, but it stops being the only tool.
Who runs this after you finish?
Your own people. The triggers sit in the tool you already use, the plays are written for the person who will run them, and the churn report is a page somebody in your team produces each month. We run the first cycle alongside you so the work is practised rather than only documented. Some businesses keep us for a quarterly review afterwards. Many do not need to, and we will say so.
What counts as churn in a business like ours?
That is the first thing we settle, because it changes every number after it. A subscription business counts a cancellation. A distributor has no cancellation at all, so churn is defined as a customer who has not ordered within a window their own history says is abnormal. A clinic may count a patient who missed a scheduled follow-up. We agree the definition with you in writing and apply it consistently across every period.
Do we need a CRM in place before starting?
No. Much of this analysis starts from an order book, a billing export or a spreadsheet, and that is enough to build a baseline and find the pattern. A CRM helps later, when triggers need to fire automatically and flags need an owner attached. If your records genuinely cannot show who bought what and when, we will tell you that before taking on the work rather than after.
Talk to us
Let us count what left last year, and name the reasons.
The assessment is a conversation, not a pitch. We will look at what your records can already tell you about churn and say plainly whether this work is what you need, or whether something simpler would do.
- No obligation and no sales script
- A reply from someone who does the work
- Your details are never sold or shared