Software Cases 2026: Churn Sets the Ceiling, in Numbers

Why the first bill is the wrong number in a software case

A client on a subscription is worth the stream of payments they make, discounted and reduced by the chance they leave. The first invoice is a fraction of that. So every question in a software case has to be answered on the stream rather than the transaction.

Here is the business to work with. A vertical software company serves 8,000 customers at $9,000 of annual revenue each, so 8,000 customers at $9,000 is $72 million of annual recurring revenue. Gross margin is 80 percent, which means 20 percent of $9,000 is $1,800 of annual cost to serve, and $7,200 of annual gross profit per customer.

Two more inputs. Monthly customer churn is 2 percent, and it costs $18,000 of sales and marketing to acquire a customer. The sales organization signs 200 new customers a month.

Those five numbers answer every question the case will ask, and they do it in a specific order. Convert the churn, get the customer life, get the lifetime value, compare it to the acquisition cost, then find the ceiling. Doing them out of order produces figures that are individually right and jointly meaningless.

The sector is growing fast enough that a client will assume the growth is available to it. The Census Bureau put US software publisher revenue at $176,958 million in the second quarter of 2026 against $154,219 million a year earlier, which is a rise of 14.7 percent, checked 30 September 2026 at census.gov. Data processing and hosting grew faster still, from $103,160 million to $124,080 million, which is 20.3 percent. Both figures are seasonally adjusted, the current quarter is preliminary and the earlier quarters are revised, so date them when you use them.

Compounding churn in a SaaS case, and the 24 percent that is 21.5

The first conversion is where most candidates lose a mark, and it takes ten seconds to get right.

Two percent monthly churn does not mean 24 percent annual churn. Survival compounds, so the fraction remaining after a year is 0.98 raised to the twelfth power. That is 0.7847, so 78.5 percent of customers remain and 21.5 percent have gone.

The naive multiplication overstates the annual loss by 2.5 percentage points, which is 11.5 percent of the true figure. On a $72 million base that gap is roughly $1.8 million of ARR, which is enough to change a recommendation about whether to hire more sales capacity.

The same compounding runs over longer horizons and the error grows. Over three years, 0.98 to the thirty sixth power is 0.4832, so 48.3 percent of a cohort survives. The multiplication route would predict 28 percent surviving, which is wrong by a factor of 1.7.

There is a second figure worth having ready. With no new sales at all, ARR halves when 0.98 to the power n equals 0.5, which happens at n of about 34 months. So this business would lose half its revenue in under three years if the sales team stopped entirely. That number makes the fragility concrete in a way a percentage does not.

Average customer life follows from the churn rate directly, on the standard assumption that the monthly rate is constant. Life is one divided by 0.02, which is 50 months, or 4.17 years. That assumption is worth stating out loud, because real churn is front loaded: new customers leave faster than old ones. Saying so and then using the constant rate anyway is the correct move in a timed case.

Lifetime value and payback in a SaaS case, and their ratio

Now the three figures a partner will ask for.

Monthly gross profit per customer is $7,200 divided by twelve, which is $600. Lifetime value on gross profit is 50 months times $600, which is $30,000. Against a $18,000 acquisition cost, the ratio is $30,000 divided by $18,000, which is 1.67.

Payback is the acquisition cost divided by monthly gross profit, so $18,000 divided by $600, which is 30 months. That is two and a half years before a customer has repaid what it cost to win, on a customer whose expected life is just over four years.

Read rows five and six together. A ratio of 1.67 on a 30 month payback describes a business that works arithmetically and cannot be scaled quickly, because each new customer consumes cash for two and a half years. Growth of that kind has to be funded, and the funding requirement rises with the growth rate.

Compare that with how a firm sets the bar in a non subscription case. ZS puts its breakeven over three years and converts it into a required market share, noting that one market "offers high $300 margins and 5% share to break even", checked 29 September 2026 at zs.com. A 30 month payback and a three year breakeven are the same kind of statement, and a candidate should ask which form the client uses before computing either.

Churn and acquisition cost enter a software case the same way

A common claim about software cases is that churn is the strongest lever. Half of that is right, and the half that is wrong is worth correcting because it comes up under questioning.

Lifetime value divided by acquisition cost is monthly gross profit, divided by the churn rate, divided by the acquisition cost. Churn and acquisition cost both sit in a denominator, so a given percentage improvement in either produces the same improvement in the ratio. Revenue per customer sits in the numerator and enters linearly.

Run the three. Cutting churn 25 percent, from 2.0 to 1.5 percent, takes life to 66.67 months and lifetime value to $40,000, a rise of 33.3 percent. Cutting acquisition cost 25 percent, from $18,000 to $13,500, takes the ratio from 1.67 to 2.22, also 33.3 percent. Raising revenue per customer 25 percent, to $11,250, lifts gross profit to $9,000 a year and lifetime value to $37,500, a rise of 25 percent.

So on the ratio alone, churn and acquisition cost are interchangeable and revenue per customer is weaker. That is the precise version of the claim, and it is the one that survives a follow up.

What makes churn genuinely different is a second effect that acquisition cost does not have. Churn also determines how large the business can get, and no amount of acquisition efficiency changes that. The next section is where that shows up, and it is the strongest argument available in this case type.

The ceiling a churn rate puts on the business

A subscription business with a constant flow of new customers and a constant churn rate reaches a plateau. The plateau sits where monthly additions equal monthly losses.

At 200 new customers a month and 2 percent monthly churn, losses equal additions when the base is 200 divided by 0.02, which is 10,000 customers. At $9,000 each that is $90 million of ARR, and the business cannot pass it while both rates hold.

The company has 8,000 customers today, so it is already at 80 percent of its own ceiling. Everything the sales team achieves from here goes into replacing departures rather than into growth, and the growth rate will decay toward zero without anyone doing anything wrong.

Now suppose the board wants $150 million of ARR. That needs $150 million divided by $9,000, which is 16,667 customers. At 2 percent churn the required addition rate is 16,667 times 0.02, which is 333 a month. Getting from 200 to 333 is a 66.7 percent increase in sales throughput, and at $18,000 a customer it costs an extra $28.7 million a year of acquisition spend.

The alternative reaches the same place without hiring. Cut monthly churn from 2.0 to 1.2 percent and the existing 200 additions support 200 divided by 0.012, which is 16,667 customers. A 40 percent reduction in churn substitutes exactly for a 66.7 percent increase in sales capacity.

That comparison is the recommendation this case wants. It names two routes to the same target, prices one of them, and puts the burden on the client to say what a 40 percent churn reduction would cost. Oliver Wyman reaches for the same lever in its own published case, recommending a subscription because it provides "consistent, recurring revenue, helping to cover fixed costs over time and ensuring financial stability", checked 30 September 2026 at oliverwyman.com.

Why 115 percent retention can sit on 21.5 percent churn

Boards see net revenue retention and read it as a statement about loyalty. It is a statement about two things at once, and the arithmetic separates them.

Gross revenue retention here is the survival rate, so 78.5 percent. Losing 21.5 percent of $72 million means 21.5 percent of $72 million is $15.48 million of ARR walking out the door each year.

Net revenue retention adds expansion from the customers who stayed. For the net figure to reach 115 percent, expansion has to cover the churn and then add 15 points. That means 36.5 percent of $72 million is $26.28 million of expansion revenue in the year. Check it: $72 million minus $15.48 million plus $26.28 million is $82.80 million, and $82.80 million divided by $72 million is 1.15.

So a 115 percent headline on a 21.5 percent logo loss requires expansion equal to more than a third of the starting base. That is a demanding number, and it means the installed base is being priced up or upsold hard. It also means the headline conceals a churn problem instead of denying one.

The practical consequence for the case is a question to ask. Request gross retention alongside net retention, and request them by cohort. A business at 115 percent net and 95 percent gross is healthy. A business at 115 percent net and 78.5 percent gross is running an expansion treadmill, and it will stop working when the surviving customers run out of modules to buy.

The loss that funds the growth, in one software case

The last step assembles a profit line, because software cases often end with a question about when the company makes money.

Gross profit is 8,000 customers at $7,200 is $57.6 million. Assume $40 million a year of research, development and general overhead, which is fixed against customer count. That leaves $17.6 million.

Acquisition spend is the sales engine. At 200 new customers a month the company signs 2,400 a year, and 2,400 new customers at $18,000 is $43.2 million. Subtract that and the annual result is $17.6 million minus $43.2 million, which is minus $25.6 million.

Read the two lines together and the business becomes legible. The installed base is profitable and the growth is expensive, and the loss is the growth rather than a symptom of a broken product. If the company stopped selling entirely it would earn $17.6 million a year and shrink 21.5 percent a year, which is a real option with a real number attached.

That framing is what an interviewer is listening for. State the profit of the base, state the cost of growth, and then say which of the two the client is actually choosing between. Altman Solon publishes "Business judgement" among the five dimensions of its business case assessment, checked 29 September 2026 at altmansolon.com, and this is the kind of statement that line is looking for.

Where a software answer loses its points

The weights below are read from our own grading code in src/lib/grading. They are ours, no consulting firm publishes one, and the reasoning for each is on how an answer is assessed.