Revenue Modeling
Revenue modeling turns vague growth goals into an explicit equation — traffic × conversion × AOV — and shows you which input you actually need to fix.
Revenue Modeling
Building a quantitative model of revenue from its underlying drivers — usually traffic × conversion rate × AOV — to make assumptions explicit.
Revenue modeling is the practice of decomposing top-line revenue into the multiplicative drivers that produce it, then assigning a numeric assumption to each. For a Shopify store the standard shape is sessions × conversion rate × average order value; for a subscription brand it extends into activation, retention, and expansion.
The value isn't the forecast number — that's almost always wrong. The value is that the model surfaces which assumption your plan is most sensitive to, and therefore where measurement, testing, and investment should concentrate next quarter.
A good revenue model is small enough to fit on one screen and explicit enough that two people looking at it disagree about specific numbers, not vibes. If your finance plan says 'we'll grow 40% next year' without naming the input that delivers it, you don't have a model — you have a target.
The format depends on the business shape. Transactional stores model per-session economics; subscription brands model cohort retention curves; marketplaces model both sides. The discipline is the same: name every lever, give it a number, and check whether the resulting total is plausible.
Revenue = Sessions × Conversion Rate × Average Order Value
Sessions
Sessions
Total store sessions in the period — paid, organic, direct, email, social combined.
CR
Conversion Rate
Share of sessions that complete a purchase, expressed as a decimal.
AOV
Average Order Value
Mean revenue per completed order, net of discounts but gross of refunds.
A mid-size apparel store on Shopify planning Q4.
Sessions (quarter): 2,400,000
Conversion rate: 2.1%
AOV: €68
→ €3,427,200 quarterly revenue
A 10% lift on CR (to 2.31%) adds ~€343k; a 10% lift on AOV via bundle tests adds the same. A 10% lift on sessions costs paid-media spend to acquire — which is why CR and AOV usually deserve the first hour of attention.
Once the model is built, run sensitivity on each input: hold the others flat and shift one by ±10%. The lever with the largest revenue swing per unit of effort is where the next experiment, redesign, or paid-media test should land.
Where uncertainty typically lives in a DTC revenue model
| Driver | Typical assumption range | Forecast error if wrong | Where to tighten it |
|---|---|---|---|
| Paid sessions | ±15-25% | High — scales linearly with spend volatility | Channel-level CAC ceilings, daily pacing |
| Organic sessions | ±10-20% | Medium — slower to move but compounds | 12-month trailing trend, seasonality index |
| Conversion rate | ±0.3-0.8pp | High — multiplies through every session | Funnel analytics, on-site experimentation |
| Average order value | ±€5-15 | Medium — bundles and shipping thresholds dominate | Cart composition, cross-sell tests |
| Repeat rate (90d) | ±5-10pp | High for subscription/replenishment SKUs | Cohort retention curves, lifecycle email |
Revenue modeling sits inside the broader practice of revenue intelligence — connecting the model's assumptions back to live behavioural data so the numbers update as the quarter progresses, rather than being re-typed into a spreadsheet every Monday.
Revenue modeling FAQ
The model is the equation — the named drivers and their relationships. The forecast is one specific run of that model with a chosen set of input assumptions. You only need one model; you'll produce many forecasts from it (base case, stretch, downside).
Granular enough that each input maps to something you can measure weekly and influence with a specific lever. Too coarse and you can't act on it; too fine (per-SKU, per-zip) and the model breaks every time a product launches. Five to twelve drivers is the typical sweet spot.
By channel, almost always. Paid social, paid search, organic, email, and direct have very different CR, AOV, and CAC profiles. An aggregate model hides which channel is actually paying for the growth and which is freeloading on brand demand.
Use a trailing 90-day blended CR as the base case, then segment by device (mobile is usually 30-50% lower than desktop) and by traffic source. If you're planning a CRO programme, bake in a conservative uplift — 5-10% on tested surfaces per quarter — rather than assuming a flat number.
For apparel and accessories, 1.5-2.5% is the typical band. Beauty and consumables run 2-3.5%. High-consideration categories (furniture, electronics over €500) sit at 0.5-1.2%. Anything you model above 4% should have specific evidence behind it.
Rebuild the structure annually — when the business shape changes (new channel, subscription launch, international expansion). Re-run the forecast monthly with fresh actuals. If you're rebuilding the equation quarterly, the model is too brittle.
Transactional models multiply per-period drivers (sessions × CR × AOV). Subscription models multiply cohort drivers (signups × activation × monthly retention) and sum across active cohorts. Replenishment brands sit between the two — model both first-order economics and 90-day repeat rate.
Assuming sessions grow linearly with ad spend. CAC rises as you scale a channel — the second €100k of Meta spend usually buys 20-40% fewer purchases than the first. Build diminishing returns into paid-traffic assumptions or your forecast will overshoot by Q3.
No. A clear spreadsheet with named drivers and one realistic scenario beats a finance-built model that nobody on the growth team can edit. Finance should pressure-test it; ownership belongs with whoever runs the channels and the site.
The model tells you which input matters most; experimentation moves that input. If CR contributes more revenue per percentage-point lift than AOV does, your test backlog should be 70% checkout and product-page work, not 70% bundle-builder ideas. Without the model, prioritisation is gut feel.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.