Reseller Hosting Client Churn Rate in Year One: A Modeled Study of 1,248 Accounts
In our modeled study of 1,248 reseller hosting accounts, 18.8% cancelled before the end of year one. That is roughly one client in five. The damage is front loaded: 38% of all first year cancellations happened in the first 90 days.
One thing before the numbers. This is a modeled dataset. We built it as a worked example of how to measure the reseller hosting client churn rate, and it is not an export of SkyNetHosting customer records. Every figure below is labeled that way, because a statistic you cannot trace back to real billing data should never be passed around as an industry benchmark.
So why publish it? Because most advice on churn stops at a single percentage, and a single percentage hides the interesting parts. When clients leave. Who leaves. What they say on the way out, and how much revenue walks out with them. Those are the questions this study works through, and you can rerun every one of them on your own billing export.
What Does First Year Churn Look Like Across 1,000+ Reseller Hosting Accounts?
Across 1,248 modeled accounts, 234 cancelled before reaching month 12. That is 18.8% churn and 81.2% retention. In plain terms, about one account in five did not make it through year one.
Churn rate is the number of customers lost in a period divided by the number you had at the start. For a hosting company, the customer here is the reseller account itself, the freelancer or agency that bought the plan from you. Their own end clients are a separate layer, and this study does not measure them. Here is the whole year in one table.
| Metric | Modeled result |
| Starting accounts | 1,248 |
| Accounts reaching 12 months | 1,014 |
| Accounts churned during year one | 234 |
| Gross logo churn | 18.8% |
| Year one retention | 81.2% |
| Gross revenue churn | 15.5% |
Total accounts included in the dataset
The dataset holds 1,248 accounts, and each one is counted exactly once, from the day its first billing term began. There are no repeat signups and no second chances in the count. That matters, because counting a reactivated account twice is one of the quickest ways to make churn look better or worse than it is.
The accounts split three ways. By acquisition quarter, 312 started in Q1, 306 in Q2, 318 in Q3 and 312 in Q4. By customer type there are 436 freelancers and solo developers, 378 small web agencies, 254 design and marketing agencies, and 180 developers and technical businesses. By billing model, 512 pay monthly, 186 quarterly, 463 annually and 87 on a two year prepaid term.
That is a decent spread, but it is still one modeled sample. In the two year prepaid group of 87, a single account moves the rate by more than a full point.
Accounts retained through the first year
Of the 1,248 accounts, 1,014 reached month 12. That is 81.2% retention, and the retained group has a median tenure of 12 months or more, which is as far as this study can see. Anything past the first anniversary sits outside the data.
Retention is the mirror image of churn, and the two always add to 100%. We report both because readers react to them differently. “18.8% churn” sounds like a problem. “81.2% retention” sounds like a result. They are the same number.
Here is the friction in that framing. Retention at month 12 says nothing about how healthy those 1,014 accounts are. A reseller who paid for twelve months and has not logged in since month four still counts as retained. Account activity, ticket volume and resource usage belong next to the retention figure, and this model does not include them.
Accounts cancelled before the first renewal
A total of 234 accounts cancelled during year one. Of those, 209 were voluntary, meaning the client chose to leave, and 25 were involuntary, meaning the account ended after a payment failure. That works out to 16.7% voluntary churn and 2.0% involuntary churn, for the 18.8% total (rounding makes the parts look a tenth short).
Another 21 accounts hit a failed payment and recovered. They are not counted as churn. The split is worth keeping because the fixes differ. Voluntary churn needs a better product or better onboarding. Involuntary churn needs better payment retries and clearer billing emails.
Then there is revenue. The 1,248 accounts started with about $25,900 in monthly recurring revenue, and the 234 churned accounts represented about $4,020 of it. That is 15.5% gross revenue churn against 18.8% logo churn. The average churned account paid about $17 a month, against about $21 across the whole cohort. You can lose almost one customer in five and lose only 15.5% of your recurring revenue, because the accounts that leave tend to be the smaller ones. Report both numbers. One alone misleads.
First year churn by customer cohort
The acquisition cohorts landed in a tight band. Here are the 90 day and 12 month figures for each quarter’s signups.
| Acquisition cohort | Starting accounts | 90 day retention | 12 month retention |
| Q1 cohort | 312 | 91.0% | 80.4% |
| Q2 cohort | 306 | 92.8% | 81.7% |
| Q3 cohort | 318 | 93.4% | 82.1% |
| Q4 cohort | 312 | 94.2% | 80.8% |
| Combined | 1,248 | 92.9% | 81.2% |
Q3 finished best at 82.1% and Q1 worst at 80.4%. That gap is 1.7 points, or about five accounts in a cohort of 312. We would not read a story into it.
Q4 is the more interesting case. It had the strongest 90 day retention at 94.2% and then slid to 80.8%, below the combined average. Strong early retention did not carry through. One modeled year cannot tell you why, so treat it as a prompt to check your own Q4 signups. Seasonal buyers, such as people signing up for a short holiday project, would be a reasonable guess to test.
Methodology and definition of a churned account
A churned account, in this study, is a reseller account that was cancelled, or ended after a payment failure, before it reached month 12. Each account is counted once. This is gross logo churn: it counts accounts lost and gives no credit for upgrades, add ons or accounts that came back later.
The 21 accounts that failed a payment and recovered stay in the retained group. The 25 that cancelled after a failed payment count as involuntary churn. The model does not break out suspended accounts as their own category, so a suspension only counts if it ended in a cancellation. If you replicate this on your own data, decide that rule before you pull a single number.
Now the disclosure that matters most. This dataset is modeled. We set the headline close to the roughly 18% median annual client account churn that ChemiCloud has reported for its reseller base, rather than picking a flattering low number, and then built the monthly, segment and reason breakdowns around it. That makes it a useful teaching example and a poor benchmark. A real export from your own billing system will beat it every time.
When Are Reseller Hosting Clients Most Likely to Cancel?
Early. Month one is the worst single month, with 38 of 1,248 accounts leaving, or 3.0%. The first three months produced 89 of the 234 cancellations, which is 38.0% of all year one churn from just a quarter of the calendar.
Early cancellations during the first few months
Month one lost 38 accounts, month two lost 27 and month three lost 24. Measured against whoever was still active at the start of each month, those are churn rates of 3.0%, 2.2% and 2.0%. After that the rate drops to 1.6% and never rises above it again.
The early leavers are a mixed group. Some bought the wrong plan. Some never finished setting up billing or moving their first client. Some were always testing, and a 30 day money back guarantee, like the one we offer, invites exactly that. If your host has a refund window, month one will look worse than the product deserves.
Still, 89 accounts in 90 days is a lot of lost effort, because you paid to acquire every one of them. The first quarter of an account’s life is where onboarding work earns the most.
Churn around the first renewal
There is a small bump at the end of the year, not a cliff. Month 10 loses 1.3% of the accounts remaining, month 11 loses 1.4% and month 12 loses 1.6%. In absolute terms that is 14, 15 and 16 cancellations. Month 12 ties the rate from months 4 and 5.
Why so mild? Billing models. Of the 1,248 accounts, 698 (55.9%) pay monthly or quarterly, so they face a cancel or continue moment many times a year. Only 463 accounts (37.1%) are on annual terms and hit a renewal decision once. The 87 on two year prepaid terms never reach one inside the window. A renewal effect spread across all those billing cycles is hard to see in a month by month line.
We cannot pull the annual renewals out of the month 12 figure with this data. That would need billing date detail the model does not carry. If you have it, look at your annual group on its own. We would expect the renewal spike to be much sharper there.
Retention after the first year
This study stops at month 12, so it cannot say what year two looks like. The 1,014 retained accounts leave the data at that point.
You can still do some arithmetic. If an annual churn rate held steady, average customer life would be roughly one divided by that rate. The Wikipedia entry on churn rate gives the same relationship: an annual rate of 25 percent implies an average customer life of four years. For 18.8%, that comes to about 5.3 years.
Churn is not steady in this dataset, though. It is highest in months one to three and settles near 1.4% a month later. A steady 1.4% a month would compound to about 15.6% a year. So if year two follows the same pattern, year two churn should come in below 18.8%. That is a hypothesis, not a result. Track the 1,014 survivors as their own cohort through months 13 to 24 before you believe it.
Account tenure and cancellation patterns
Here is the full month by month schedule.
| Month | Starting accounts | Churned that month | Monthly churn | Remaining |
| 1 | 1,248 | 38 | 3.0% | 1,210 |
| 2 | 1,210 | 27 | 2.2% | 1,183 |
| 3 | 1,183 | 24 | 2.0% | 1,159 |
| 4 | 1,159 | 19 | 1.6% | 1,140 |
| 5 | 1,140 | 18 | 1.6% | 1,122 |
| 6 | 1,122 | 17 | 1.5% | 1,105 |
| 7 | 1,105 | 16 | 1.4% | 1,089 |
| 8 | 1,089 | 15 | 1.4% | 1,074 |
| 9 | 1,074 | 15 | 1.4% | 1,059 |
| 10 | 1,059 | 14 | 1.3% | 1,045 |
| 11 | 1,045 | 15 | 1.4% | 1,030 |
| 12 | 1,030 | 16 | 1.6% | 1,014 |
| Total | 234 | 1,014 |
Two patterns stand out. First, months 1 to 3 averaged about 30 cancellations a month and months 4 to 12 averaged about 16, so the early pace is nearly double the later one.
Second, by the end of month 5, 126 of the 234 cancellations had already happened. More than half of all year one churn is finished within five months. If you only have capacity for one retention project, point it at those five months.
Retention curves across different customer cohorts
Plot the share of accounts still active and you get a curve that drops fast and then flattens. From the monthly schedule, 92.9% of accounts remain at the end of month 3, 88.5% at month 6, 84.9% at month 9 and 81.2% at month 12. The first three months cost 7.1 points. Each three month stretch after that costs between 3.5 and 4.5.
Cut by billing model, the curves separate hard. At month 12, 74.4% of monthly accounts remain, against 79.6% of quarterly, 86.8% of annual and 95.4% of two year prepaid accounts. Those are year end snapshots only. The model does not give month by month curves for each segment.
The gaps reflect commitment length as much as satisfaction, a point we return to under pricing. Build the same curves for your own accounts by signup quarter and by billing model. Two cuts, and you will see more than any single churn percentage shows.
Why Do Reseller Hosting Clients Leave During Their First Year?
Mostly for reasons tied to the hosting product, its price or the provider. In the modeled data, 171 of the 234 cancellations, or 73.1%, came from price, scalability, performance, support, competitor moves, billing problems or security concerns. Another 60, or 25.6%, came from a client’s business closing or its client portfolio shrinking. Three accounts, 1.3%, gave other reasons.
| Cancellation reason | Accounts | Share of churn |
| Business or project closed | 42 | 17.9% |
| Found cheaper hosting | 39 | 16.7% |
| Outgrew reseller plan | 34 | 14.5% |
| Performance or resource limitations | 31 | 13.2% |
| Poor or slow support | 27 | 11.5% |
| Moved to competitor | 25 | 10.7% |
| Client portfolio declined | 18 | 7.7% |
| Billing or payment problems | 10 | 4.3% |
| Security concerns | 5 | 2.1% |
| Other | 3 | 1.3% |
| Total | 234 | 100% |
Read “tied to the provider” carefully. It is not the same as the provider’s fault. Outgrowing a plan is a growth signal, and finding cheaper hosting is as much about a reseller’s margins as it is about service. Even so, the old assumption that clients leave only because the hosting is bad does not survive these numbers. A quarter of them left for reasons no host could have fixed.
Pricing and renewal costs
Price shows up twice. Thirty nine accounts, 16.7% of all churn, left for cheaper hosting. The price band table shows the same thing from the other side.
| Monthly price | Accounts | First year churn |
| Under $5 | 174 | 29.3% |
| $5 to $9.99 | 296 | 24.0% |
| $10 to $19.99 | 407 | 18.2% |
| $20 to $49.99 | 279 | 12.9% |
| $50 and up | 92 | 9.8% |
Accounts paying under $5 a month churned at 29.3%. Accounts paying $50 or more churned at 9.8%. That is roughly a threefold gap. The likeliest reading is that switching costs are tiny at $5 a month, and the bigger the bill, the more thought goes into a move.
Billing model matters nearly as much.
| Billing model | Accounts | Churned | First year churn |
| Monthly | 512 | 131 | 25.6% |
| Quarterly | 186 | 38 | 20.4% |
| Annual | 463 | 61 | 13.2% |
| Two year prepaid | 87 | 4 | 4.6% |
| Total | 1,248 | 234 | 18.8% |
Be careful reading that. A client who commits for a year is not automatically a happier client. They simply have fewer chances to cancel. Billing period churn and dissatisfaction are different things.
Our terms ask for five days notice before the anniversary billing date to cancel a month to month service, so a monthly client can leave quickly, and we do not pretend otherwise. For resellers selling to freelancers on thin margins, our Budget Reseller Plans are the low cost entry point. They suit a handful of client sites. They are not the plan to be on once you manage fifty accounts.
Performance and resource limitations
Thirty one accounts, 13.2% of churn, left over performance or resource limits. That group is the most fixable one on a host’s side, because the cause sits on the server and not in the client’s head.
Limits bite in places people forget to watch. On our Budget Reseller Plans the terms set 50,000 inodes per cPanel user and 250,000 across the reseller account. Every file, email and image uses one inode, so a client with a photo heavy site can hit the per user cap long before disk space or bandwidth looks alarming. The site then misbehaves in confusing ways.
Hardware helps. Our servers run Intel Dual Xeon processors with NVMe drives, and we use LiteSpeed in place of Apache. None of that rescues an account that is packed too full. We tell resellers to stop filling any limit at 80%, so a plan allowing 50 accounts gets 40, which leaves room for the client who uploads a huge gallery overnight. For the full list of what each plan includes, you can compare our reseller features.
Support and service experience
Support and service experience pushed out 27 accounts, 11.5% of cancellations. Slow answers during setup are the usual culprit. A reseller who waits a day for help with a first client migration is already shopping.
Of all the causes in the table, this is the one a host controls most directly, and the one clients remember longest. A reseller who gets a real answer at 2 a.m. forgives a lot. We run 24/7 live support for that reason, and our End User Support option puts our team in front of your own clients’ questions so they do not all land on you.
Here is the honest limit. Support only retains clients who stay long enough to need it. Someone who leaves in week three because the plan was wrong never meets your best technician.
Business closure or declining client demand
A client’s business shutting down, or its client portfolio shrinking, ended 60 accounts. That is 42 closures (17.9% of churn) and 18 declining portfolios (7.7%), 25.6% in total. A quarter of year one churn had nothing to do with hosting quality.
That changes what you can fix. You cannot retain a freelancer who took a full time job. You can avoid being the host that keeps billing them for servers they no longer need. Freelancers are the segment most exposed here, as the customer type data below shows.
For resellers who sell VPS or dedicated servers on to their own clients, our Resell VPS and Servers option, open to VIP and Corporate VIP clients, carries an extended cancellation policy. You only renew those services if your own client renews with you. That is a direct answer to the closure problem, though it applies to resold servers and not to every plan.
Moving to another hosting provider
Twenty five accounts moved to a named competitor, 10.7% of churn. Add the 39 that left for cheaper hosting and switching of one kind or another explains 64 cancellations, or 27.4%.
Treat those reasons with some suspicion. Cancellation forms collect what people choose to say, and “cheaper elsewhere” is the easiest thing to say. Behind it could sit a support delay, a slow server or a billing surprise. The model’s “other” category holds only 3 accounts, which suggests the reasons were tidy. Real exit data rarely is.
A short follow up message a week after cancellation often gets a more honest answer than the form does. Ask one question. Which provider did you move to, and what made the difference? Competitor names tell you who you actually lose to, which is rarely who you assume.
Outgrowing the original reseller plan
In all, 34 accounts, 14.5% of churn, left because the reseller plan was too small. In a sense this is the happy kind of churn, because the client’s business grew. It only stays happy if they grow into your next product instead of someone else’s.
The pattern we see most is a hybrid. A reseller keeps the small client sites on the reseller plan and moves the two or three heaviest onto a separate VPS. Costs stay low for the small clients, and the big ones stop competing for shared limits. When a reseller needs full control of the hardware for a large agency workload, dedicated servers are the next step. And a reseller who wants to sell hosting to other resellers should look at Master Reseller Hosting.
In this dataset those 34 accounts left, though some may have moved up elsewhere. The point stands. An upgrade path you cannot see is an upgrade path you will not offer in time.
How Does First Year Churn Differ Between Different Types of Resellers?
Freelancers and solo developers churn most, at 23.9%. Agencies churn least, at 15.0% to 15.1%. Developers and technical businesses sit between them at 19.4%.
| Customer type | Accounts | Churned | First year churn |
| Freelancers and solo developers | 436 | 104 | 23.9% |
| Small web agencies | 378 | 57 | 15.1% |
| Design and marketing agencies | 254 | 38 | 15.0% |
| Developers and technical businesses | 180 | 35 | 19.4% |
| Total | 1,248 | 234 | 18.8% |
Freelance web designers
Freelancers and solo developers make up 436 of the 1,248 accounts, about 35%, and 104 of the 234 cancellations, about 44%. Their first year churn rate is 23.9%. Nearly a quarter leave.
A freelancer’s hosting portfolio rides on a handful of end clients. Lose two projects and the reseller plan stops paying for itself. A career change, a full time job offer or a client who goes quiet all hit a one person business harder than a team. That matches the closure numbers above.
One caution about the label. This segment mixes freelance web designers with solo developers because the model groups them. If you can separate the two in your own data, do. They may behave differently.
Digital and web agencies
Small web agencies churned at 15.1% (57 of 378) and design and marketing agencies at 15.0% (38 of 254). Put together, 95 of 632 agency accounts left, which is 15.0%. That sits almost nine points below the freelancer rate.
Agencies have several clients, ongoing retainers and someone responsible for hosting. Moving every client site to a new host is real work, and an agency will not do it over a small price difference. They are not immune to leaving. They need a bigger reason. Most of the agency effect is probably about client count and switching effort, but the model cannot prove that.
Developers and technical resellers
Developers and technical businesses sit at 19.4%, with 35 of 180 accounts leaving. They are comfortable with migrations, DNS and server tooling, so the effort of leaving is low for them. They also notice resource limits and performance faster than a designer would.
Our read is that they churn for reasons closer to the performance and outgrowing categories, and less for closure. The model does not tag cancellation reasons by segment, so that is a reading and not a result. If you can cut your own cancellation reasons by customer type, this is the first cut we would make.
New hosting businesses
The model has no separate segment for brand new hosting businesses, so we will not invent a number. What we can say is what to check. A new host has no client base yet, which means its first year is a race between landing clients and paying for the plan. It is reasonable to expect them to look closer to freelancers than to agencies, though that is an expectation and not a measurement.
If you sell to people starting hosting companies, tag them at signup. Then track whether each has added a first paying end client by day 30, 60 and 90. That one milestone is a better early warning than any satisfaction survey. A free WHMCS license, which we bundle with reseller plans at a $15.95 a month value, exists to take billing off the list of things a new host must build before earning a first dollar.
Established businesses adding reseller hosting
The model also lacks a segment for established businesses that add reseller hosting to an existing offer, such as an IT firm or a print shop. No number here either. Their situation differs. They already have customers, billing and support habits, and reseller hosting is a side product they can keep running even if it is not growing.
That can make their churn lower and slower. It can also make them quietly neglectful, with accounts that persist and never generate value. Track active usage next to retention for this group. A retained account that nobody logs into is a future cancellation.
How Does Reseller Hosting Churn Compare With Other Hosting Benchmarks?
The closest comparison is another hosting specific figure. ChemiCloud has reported a median of about 18% annual client account churn across its reseller base, and our modeled 18.8% sits right next to it. Do not read that as confirmation. We set the model near that figure on purpose, so the match proves nothing.
Differences between hosting and general subscription businesses
Industry matters more than almost any other variable. CustomerGauge’s 2025 research across 11 B2B industries found average churn running from 11% in energy to 56% in wholesale, which you can see in their churn rate by industry breakdown. A spread that wide tells you a single “subscription churn rate” is not a thing.
Hosting is a subscription business with traits most software subscriptions lack. Clients depend on DNS, email and live websites, which makes switching a chore. At low price points, though, the chore is small enough to finish over a weekend. That is why the price band table earlier shows such a steep slope.
Annual versus monthly churn measurements
Annual churn equals one minus the quantity (one minus the monthly rate) raised to the 12th power. Wall Street Prep lays out that conversion in its churn rate guide. A flat 1.5% monthly rate looks small, yet it compounds to 16.6% over a year.
Monthly figures also mislead when churn is uneven. In this dataset month one ran at 3.0% and month ten at 1.3%. Quote either one alone and you describe a different business. The cohort approach matters for the same reason. Wikipedia notes that measuring churn across a fast growing aggregate base understates the true rate compared with following a cohort, because the denominator is padded with accounts that have not had time to leave.
Why customer type and pricing affect churn
Customer type and price do most of the work in this dataset. Freelancers churn at 23.9% and agencies at 15.0%. Accounts under $5 a month churn at 29.3% and accounts at $50 or more churn at 9.8%. Both gaps are far wider than any difference between acquisition quarters.
That has a practical effect on benchmarking. Imagine two hosts. One sells almost entirely to agencies on mid priced plans, and the other sells budget plans to freelancers. Mix alone could put their year one churn nearly nine points apart, even with identical service quality, because nine points is the gap between the freelancer and agency rates here.
So a host comparing its overall churn rate with someone else’s is often comparing customer mix and not service quality. Compare like with like, or do not compare.
The limitations of applying SaaS churn benchmarks to hosting
Software as a service benchmarks are the ones most often borrowed, and they transfer poorly. A SaaS subscription is usually priced per seat or per tier and can grow through expansion revenue, which offsets churn in the net numbers. A reseller hosting account is infrastructure. It carries domains, DNS, email and live websites, so leaving has a migration cost that most software subscriptions do not have.
Hosting also mixes billing terms inside one customer base. In this dataset 512 accounts paid monthly while 87 sat on two year prepaid terms, and their churn ran from 25.6% down to 4.6%. A single blended benchmark flattens that.
We are not saying SaaS figures are useless. As a sanity check on direction they are fine. As a number to hit, they are the wrong yardstick.
Using external benchmarks as context rather than a target
Use outside figures to ask a better question, not to set a target. If ChemiCloud’s reported 18% is roughly where a comparable host sits, that tells you a 30% rate of your own deserves a look. It does not tell you 18% is good, because the right number depends on your customer mix, price points and billing terms.
A target worth setting is internal. Pick your highest churn segment and bring it down a few points over the next four quarters. Then watch whether the first 90 days improve. Beating your own last year is a more honest scoreboard than beating a figure from another company’s customer base.
And if you cite this study, cite it as a modeled example. Please do not quote 18.8% as a hosting industry average. It is not one.
What Can Reseller Hosting Businesses Learn From the First Year Churn Data?
Act in the first 90 days, separate the causes you control from the ones you do not, and measure by cohort and reason. In the model, 38.0% of churn landed in the first 90 days, 73.1% tied to the product, price or provider, and 25.6% tied to a client’s business changing.
Identifying the highest risk customer periods
Start with the calendar of an account, not the calendar year. Months one to three produced 38.0% of year one churn, and more than half had happened by the end of month five. Month 12 shows a smaller second bump.
Put those windows into your process. Someone, or something, should check in at day 7, day 30 and day 60, when the early decisions get made. Add a renewal reminder well ahead of the anniversary for annual accounts, and a billing health check for monthly ones. This costs far less than winning back an account that has already left. Our 30 day money back guarantee makes the first month a trial period by design, so build for it.
Separating provider related churn from business related churn
Roughly 73.1% of cancellations tie to the hosting product, price or provider, and 25.6% to a client’s business changing. Do not lump them together. A host that treats a freelancer’s closed business as a service failure will chase the wrong fix. One that blames every cancellation on clients going out of business will never improve.
Give your cancellation form two main buckets and a small third for odd cases. Then divide the work. Business related churn gets flexible options such as pauses and downgrades. Provider related churn gets operational attention: support response times, server performance, plan limits and clear pricing.
Be careful with outgrowing the plan. It sits in the provider bucket in this model, and it is partly a success signal. Count it, but read it differently.
Improving onboarding before the first renewal
The first 90 days are the onboarding window, and the data says that is where the biggest single block of churn lives. A new reseller needs to reach a first paying end client quickly. Everything you can do to shorten that path helps retention.
For a SkyNetHosting reseller that means using the tools already included. A free WHMCS license for billing and provisioning, a free domain reseller account for selling domains under their own brand, and a storefront so clients can order without waiting on the reseller. Having these set up in week one beats having them available in month four.
We will say plainly that tooling does not replace selling. A reseller still has to find their own customers. No host can do that for them, and a hosting company that implies otherwise is setting up a month three cancellation.
Matching hosting resources with customer growth
Outgrowing and performance together account for 65 cancellations, 27.8% of churn, which is more than business closure. Both are visible before they happen if you look.
Review resource use on a schedule. Each quarter, list the accounts using more than 80% of any limit, whether that is disk, bandwidth, inodes or cPanel account count. Those are the clients about to hit a wall. Contact them first with a plan for the next step, before they open an angry ticket.
The next step is not always a bigger version of the same plan. Sometimes it means moving the heaviest client to its own VPS and leaving the rest where they are. Sometimes it means nothing, because the problem is one badly configured site. Working out which one it is shows the reseller you are paying attention, which is a retention tool all by itself.
Monitoring churn by customer cohort and reason
Track four cuts every quarter: signup cohort, customer type, billing model and cancellation reason. This study’s tables show why. The overall 18.8% hides a 9.8% group and a 29.3% group, and a 4.6% group and a 25.6% group.
Keep gross logo churn and gross revenue churn side by side. In the model the gap was 18.8% against 15.5%, and the gap itself says something: the accounts leaving are smaller than average. If revenue churn ever rises above logo churn, your larger accounts have started to leave, which is a far more serious situation.
Write the churn definition down once and never change it quietly. Whether suspensions count. Whether reactivations reset the clock. Whether a failed payment that recovers counts at all. A definition that drifts makes every trend line meaningless.
Building retention strategies around actual cancellation data
After more than 20 years and 700,000+ websites hosted, our view is simple. Retention work that starts from guesses wastes effort, and retention work that starts from your own cancellation data does not.
Take the modeled reasons as a template for the questions to ask, not as answers. Rank your own cancellation reasons by count and by revenue. Fix the top controllable one first. If price leads, look at billing terms and plan fit before cutting prices. If support leads, look at first response time during setup. If outgrowing leads, build the upgrade path.
Then measure again on the next cohort. One quarter of data will not prove a fix worked, and 1,248 modeled accounts will not prove anything about your business. Your own export will. Pull it, apply a written definition, and compare.
Building a hosting business and want fewer avoidable exits in the first 90 days? Start with our reseller hosting plans, which include a free WHMCS license, a free domain reseller account and 24/7 live support.
