Your CAC Payback Is a Lie the Moment You Blend Two Motions

Blended CAC payback: 17 months. The board nods, it sits under the 18-month efficiency line everyone treats as gospel, the business reads healthy. Then you split it by motion and find self-serve paying back in 6 months and enterprise paying back in 24. The blended 17 describes neither. It is the average of a sprinter and a marathoner, and if you manage to the average you will tell the sprinter to slow down and the marathoner to sprint.

For years the operating rule had one form: get blended CAC payback under the efficiency line and the go-to-market engine is tuned. That rule breaks the instant a company runs more than one motion, because the motions do not sit near each other on the number line. They sit an order of magnitude apart. A self-serve customer costs a few hundred dollars to acquire and pays back in a quarter or two. An enterprise customer costs six figures to acquire and pays back over two years. Blending them is not smoothing noise. It is fusing two different businesses into a single figure that no one can act on.

The reason the blend fails is scale. Acquisition cost per customer does not vary by 20% across your motions. It varies by 10x, then 10x again. Watch the field grow. The dot in the center is one dollar of acquisition cost, held constant, while the field around it expands to the cost of landing one customer in each motion.

1x~$300 CACPLGSelf-serve signup, no human in the loop, credit card close

−

PLGMMENT

+

one dollar of acquisition cost stays constant. The field around it grows each level.

Payback is that same field divided by how fast each customer returns gross margin. When the cost of the sale climbs by a factor of fifty and then ten again, the time to earn it back climbs with it. That is why one payback number cannot cover the whole company. It is describing a field that grows by orders of magnitude with a single point. Below is the framework that fixes it: a four-step build that tags every dollar to its motion at the source, so two honest numbers come out of the system on the same cadence as every other board metric. Locate your own reporting on it. Most teams sit at step zero, typing one blended figure into a slide.

Two motions, two clocks, one misleading average

Self-serve and enterprise are not the same business wearing one logo. They acquire differently, cost differently, and pay back on different horizons. Both can be healthy. The average of the two is healthy for no one.

Motion	New ARR	S&M cost	Gross margin	Payback
Self-serve	$3,000,000	$1,200,000	85%	5.6 months
Enterprise	$6,000,000	$9,000,000	80%	22.5 months
Blended	$9,000,000	$10,200,000	81%	16.8 months

The math on each row is S&M spend divided by new ARR times gross margin, annualized to months.

self-serve:  1,200,000 / (3,000,000 x 0.85) = 0.47 yr = 5.6 mo
enterprise:  9,000,000 / (6,000,000 x 0.80) = 1.88 yr = 22.5 mo
blended:    10,200,000 / (9,000,000 x 0.81) = 1.40 yr = 16.8 mo

Both motions are fine. Self-serve at 5.6 months is a machine you should feed more capital. Enterprise at 22.5 months is normal for large deals with long cycles and high retention, and it sits above the 18-month line, which reads as a problem only if you forget it is a different motion. The blended 16.8 tells you to do nothing, which is the wrong answer for both. Notice the two motions sit a full order of magnitude apart on the zoom above, and the payback table lands them 17 months apart in time. Same divergence, two views.

The two payback clocks, drawn against the 18-month line so the distance is visible:

View as table

Item	Value
Self-serve	5.6 mo
Blended	16.8 mo
Enterprise	22.5 mo

The blend defunds your best motion

Here is the operational damage. A board that sees 16.8 months blended, just under the line, concludes the machine is tuned and holds spend flat. But self-serve at 5.6 months is starving for capital it could turn into ARR in half a year, and enterprise at 22.5 is quietly dragging the blend toward the danger line as it grows.

Split the metric and the decision inverts. Fund self-serve harder because it pays back in under half a year. Hold enterprise steady and judge it against enterprise benchmarks, not the company average. Neither of those decisions is visible in the blended number.

The reason the blend is so dangerous is that mix is always moving. A company rarely holds a fixed ratio of self-serve to enterprise ARR. Land a few large enterprise logos and the enterprise share of new ARR jumps, which drags the blended payback out by months even though every individual deal paid back exactly as expected. A leader watching only the blend reads that as the engine getting less efficient and reaches for the brake. The engine did not change. The mix did. You cannot tell those two apart from a single number, and telling them apart is the entire point of the metric.

There is a second trap worth naming: payback and retention are linked, and the blend hides that link too. Enterprise pays back slower but retains far longer. The 2026 Aleph and Benchmarkit study (n=230) put median net revenue retention at 102%, but with a 10-point spread between usage-priced books at 108% and seat-priced books at 98%, and a gap by size from 94% under $5M ARR to 103% above $100M. Enterprise motions cluster at the top of that range. So a 24-month enterprise payback against a five-year customer life is a far better lifetime return than a 6-month self-serve payback against an 18-month life. Read payback next to gross retention by motion, never alone, or you will conclude the fast-payback motion is the more valuable one when the slow one may be worth several times more over its life.

	Managing to the blend	Managing by motion
Self-serve (5.6mo)	Held flat, treated as average	Funded harder, fastest payback in the book
Enterprise (22.5mo)	Looks like the problem, gets cut	Held steady, judged vs enterprise history
When enterprise grows	Blend worsens to 19.5mo, board panics	Expected, read as mix not decay
Net effect	Starve the sprinter, punish the marathoner	Capital flows to the return it earns

Why one number cannot describe two curves

Payback is a curve, not a point. Self-serve recovers its cost fast and flattens. Enterprise recovers slowly, then keeps paying for years because the retention is higher. Averaging the two curves into one number throws away the shape that carries the decision.

Put the levers in your own hands and watch the number move. Drop gross margin or lift spend and the payback stretches fast. This is the single-motion formula, run it once per motion:

Payback for one motion: S&M, new ARR, gross margin

S&M spend÷( new ARR×gross margin )=10

months to payback

S&M spend / period ($)$600kNew ARR won ($)$900kGross margin80.0%

TryEfficient (10mo)Bloated (28mo)

Under 12 months is efficient, 12 to 18 is normal for mid-market, past 24 months you are buying revenue faster than it pays you back. Watch this before you scale spend.

months to payback: 10

Here’s how I’d build it: payback by motion

This is the build I stand up so two honest numbers come out of the system on the same cadence as every other board metric. Four steps, in order, and the whole thing lives or dies on step one: if the dollar is not tagged to a motion at the source, everything downstream is a reconstruction you will argue about.

1. 1

   1. Tag ARR by motion on the opportunity

   Add a Motion field (self-serve, mid-market, enterprise) that is required at close. This is the join key for everything downstream. Backfill closed-won for the trailing four quarters so you have history on day one. This is the PLG, MM, and ENT split from the zoom, made into a real field.

2. 2

   2. Allocate S&M to motion in the GL

   Map each S&M cost center to a motion, even coarsely. Product and paid acquisition to self-serve. AE, SE, and field marketing to enterprise. Shared costs split on a defensible ratio. A rough allocation you disclose beats a blend that hides the split.

3. 3

   3. Write one payback formula per motion

   Payback months = (motion S&M / (motion new ARR times motion gross margin)) times 12. Save it as a metric per motion so nobody recomputes it and nobody quietly changes the margin assumption between board decks.

4. 4

   4. Report two numbers against two benchmarks

   Self-serve against the 6-to-9-month band, enterprise against the 24-month band. When a motion drifts, you see it against its own history and know whether it is a real problem or just mix shift.

Once spend is allocated, payback by motion is a two-line query refreshed on the same schedule as every other board metric. The board sees two numbers with two benchmarks instead of one number that hides the story. When enterprise payback drifts, you see it against enterprise history and know whether it is a real problem or a mix effect. When self-serve payback holds under 6 months, you have the evidence to ask for more budget with a straight face.

The allocation is where most teams stall, so make it boring on purpose. You do not need activity-based costing down to the individual campaign. You need every S&M dollar assigned to exactly one motion by a rule you can write on a single line. Here is the rule I use, and I apply it the same way every quarter:

# s&m allocation rule, applied identically each quarter
self_serve:
  - paid_search
  - paid_social
  - product_led_growth_spend
  - lifecycle_email
enterprise:
  - ae_fully_loaded_cost
  - se_fully_loaded_cost
  - field_marketing
  - events                 # unless explicitly a self-serve acquisition play
shared_split_on_arr_ratio:  # brand, revops, marketing ops
  - brand
  - revops
  - marketing_ops
gross_margin:
  rule: recompute_per_motion   # never borrow the company blended margin
  self_serve: 0.85             # no human delivery cost
  enterprise: 0.80             # carries support and services load

Write the rule down, apply it the same way every quarter, and the trend line stays honest even if the absolute number is approximate. A consistent rough allocation beats a precise one that changes definition every time someone new touches the model, because the whole value of the metric is the quarter-over-quarter movement, not the third decimal place.

One guardrail hides inside that config: recompute gross margin per motion, do not borrow the company blended margin. Self-serve usually carries higher gross margin because it has no human delivery cost, and enterprise often carries more support and services load. If you apply one blended margin to both, you have re-introduced a blend inside the very metric you split, and the enterprise payback will read better than it truly is. The margin belongs to the motion, same as the spend and the ARR.

The blended number is the easiest to compute and the most expensive to manage by. Split it once, and the capital-allocation conversation stops being an argument about a single number and starts being two conversations about two motions, each judged against the benchmark it belongs to. The expansion economics reinforce which motion to feed: at $0.80 to expand a dollar of ARR against $1.63 to acquire one (Aleph and Benchmarkit, 2026), the motion with the higher retention compounds harder every year you hold it. For the retention side of unit economics, see why NRR is the only number that compounds, and for the model that turns these motion splits into a plan you can defend, see driver-based planning.