How to use Growth Model Templates
Calibrate four DTC growth models — paid acquisition, subscription, repeat-purchase, and viral — to find the lever with the most impact on your 24-month trajectory.
Growth Model Templates
Quantitative spreadsheets that project revenue from a handful of inputs — CAC, retention, AOV, referral rate — so you can stress-test which lever matters most.
Growth model templates are pre-built quantitative frameworks that translate the levers of your business — paid spend, retention curves, average order value, referral coefficient — into a 12- to 24-month revenue projection. Instead of arguing about strategy in abstract terms, you calibrate the model with last quarter's numbers and watch what happens when you nudge one input.
Four archetypes cover most online retail businesses: paid-acquisition-driven (you buy growth), subscription (you compound MRR), repeat-purchase-driven (you compound LTV through reorders), and viral (each customer brings fractional new ones). The right template is the one whose dominant assumption matches your actual customer behaviour.
Most growth plans fail the same way: a deck full of initiatives with no shared math underneath. Marketing wants more paid budget, retention wants email spend, product wants a referral programme — and nobody can say which of those moves the 18-month number more.
A growth model fixes that by forcing one shared spreadsheet. Once your CAC, retention curve, and AOV are in the same file, the argument shifts from opinion to sensitivity analysis. The lever that bends the curve hardest is the one you fund first.
The four growth model archetypes
The paid-acquisition model assumes new revenue is roughly proportional to paid spend, throttled by blended CAC and the payback period. It fits apparel, beauty launches, and any brand where Meta and Google together drive 60%+ of new customers. The dominant lever is CAC efficiency; the danger is that a 20% CAC inflation can wipe out a year of margin gains.
The subscription model compounds MRR from a fixed acquisition cohort each month, with churn as the leak. It fits replenishment categories — coffee, supplements, pet food. Here the dominant lever is churn: a drop from 8% to 6% monthly churn extends average customer lifetime from 12.5 to 16.7 months, a 34% LTV bump with zero extra acquisition.
The repeat-purchase model is the most common DTC pattern: one-time purchases with a probability distribution of reorders over 12-24 months. AOV, reorder rate, and time-between-orders are the three inputs that matter. The viral model adds a referral coefficient (k-factor) — fractional, almost always under 0.5 — that multiplies each acquired cohort. Pure viral is rare; viral as an amplifier on a paid base is realistic.
Don't pick the model you wish you were
Brands often calibrate a subscription model when only 12% of customers actually subscribe, or a viral model when their k-factor is 0.05. The model has to match observed behaviour from your last two quarters of data, not the business you'd like to be running by Q4.
How each model trajectories over 24 months
Below is a side-by-side simulation of four brands all starting at €100k in monthly revenue, each running a different growth model with realistic inputs for a €3-8M DTC store. Same starting point, very different endings — because the underlying math compounds differently.
Notice the subscription line: it starts slowest but the compounding overtakes paid by month 18. The viral-amplified line (a paid base with k=0.25) outperforms pure paid by ~30% at month 24, which is why even modest referral programmes are worth modelling explicitly rather than ignoring.
Projected monthly revenue (€k) — 24 months, same €100k starting point
Paid-acquisition
Subscription
Repeat-purchase
Paid + viral (k=0.25)
The takeaway isn't that subscription "wins" — it's that the dominant lever is different in each model, so the same €50k of investment produces wildly different returns depending on where you put it. Modelling first, then deciding, is the whole point.
Inputs and realistic ranges for each model
Each template needs four to six inputs to produce a credible projection. Get these from GA4, your Shopify reports, and your ad platforms; if you've connected historical GA4 to your analytics, the calibration is a single afternoon's work rather than a week of data wrangling.
The ranges below reflect what's typical for online retailers in the €1M-€15M revenue band. Your numbers will sit somewhere in here; if they don't, that's the first signal worth investigating before you trust any forecast.
Typical input ranges by growth model archetype
| Model | Primary input | Typical range | Secondary input | Typical range |
|---|---|---|---|---|
| Paid-acquisition | Blended CAC | €18 – €55 | Payback period | 2.5 – 6 months |
| Subscription | Monthly churn | 4% – 9% | MRR per subscriber | €22 – €65 |
| Repeat-purchase | 12-mo reorder rate | 25% – 55% | Time between orders | 55 – 110 days |
| Viral (amplifier) | Referral k-factor | 0.05 – 0.35 | Referral conversion | 8% – 22% |
| All models | Gross margin | 55% – 72% | AOV | €48 – €120 |
Use the secondary inputs as sensitivity dials. The most useful question a model answers isn't "what's our revenue in month 18?" — it's "if we improve reorder rate from 32% to 38%, what does that do to the trajectory versus dropping CAC by €4?". Both are roughly equivalent effort; only one of them moves the number more.
Calibrating the model with your own data
Pull a clean 90-day window — ideally one without a big sale event distorting the averages — and compute each input from raw data, not from the dashboards you already trust. Dashboards round, segment, and exclude in ways that hide assumptions; the model needs the underlying numbers.
Then back-test: run the model against the previous quarter's starting state and check how close its month-3 prediction lands to actual. If it's within 10% you have a usable forecasting tool. If it's off by 25%+, the wrong archetype is fitted and you need to switch templates — usually from paid to repeat-purchase, or from subscription to a hybrid.
What a calibrated model actually changes
Teams that run quarterly model reviews stop debating tactics in isolation. The conversation becomes "this initiative moves input X by Y%, which moves revenue by Z" — and projects that don't move any input meaningfully get cut. That's the real ROI of templates: they kill the bad ideas before they consume a sprint.
Frequently asked questions
A financial forecast projects P&L line items from top-down assumptions ("we'll grow 30% next year"). A growth model builds the revenue line from the bottom up — from CAC, retention, AOV — so you can see which behaviour change drives the growth, not just that growth happens.
Run the repeat-purchase template first. It fits the majority of online stores under €10M and only needs four inputs: AOV, gross margin, 12-month reorder rate, and acquisition cost. If reorder rate is above 50% or you have a true subscription product, switch to the subscription model afterwards.
Quarterly is the sweet spot. CAC shifts with ad auction dynamics, reorder rate shifts with product mix, and a model running on six-month-old numbers will mislead you. Lock a calendar reminder for the second week of each quarter.
Yes, and you usually should. The most realistic setup is repeat-purchase as the base layer with viral as an amplifier (k-factor multiplied onto each acquired cohort). Pure single-model brands are rare above €3M revenue.
Using blended numbers that hide cohort effects. Your 12-month reorder rate looks like 35%, but if new customers reorder at 22% and customers from your top channel reorder at 48%, the average is meaningless for forecasting. Segment by acquisition channel at minimum.
No. A spreadsheet with inputs from GA4, Shopify, and your ad platforms is enough. The value is in the structure of the model, not the infrastructure. Connect a warehouse later if you want monthly auto-refresh.
Realistically, ±20-30% at month 24 if the inputs are well-calibrated. The point isn't pinpoint accuracy — it's relative comparison of scenarios. "Improving retention beats lowering CAC by 2x" is true regardless of whether the absolute number is off by a quarter.
The model tells you which input has the highest leverage; experimentation tells you whether you can actually move that input. Pair them: model says reorder rate matters most, then run tests on post-purchase email, replenishment reminders, and second-order discounting to lift it.
Organic word-of-mouth produces a k-factor too — it's just invisible without measurement. Survey new customers ("how did you hear about us?") and treat the resulting referral rate as your baseline k. Even k=0.08 compounds meaningfully over 24 months.
Apply a monthly seasonality index on top of the model output rather than baking it into the inputs. December at 1.4x, January at 0.7x, etc. This keeps the model's underlying mechanics clean and makes year-over-year comparison straightforward.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.