Unit Economics as a Marketing Decision Input

By ScaleGrowth Digital Editorial

Most marketing budgets get sized off a top-down rule of thumb (a percentage of revenue, a multiple of last quarter, a competitive benchmark) and only then bottom-up reconciled against the unit economics they need to honour. The order is wrong. Unit economics belong before the budget envelope, not after it. The right per-unit revenue number can change a quarterly recommendation by an order of magnitude. The wrong one (often a founder-stated figure that has not been pulled from the database) produces a spend plan the business cannot actually afford. This piece sets out the discipline ScaleGrowth Digital uses on every multi-location, multi-product or location-based engagement before any envelope is written.

The 8x correction that started this rule

A multi-location F&B brand with 86 active stores asked for a Q2 social, paid, and Google Business Profile strategy. The founder-stated baseline was four lakh rupees per store per month. A proposed Q2 spend envelope was drafted at 51 to 78 lakh rupees against that figure. Before shipping, a database pull from the Laravel command centre surfaced the real number across four known pilot stores: 1.58 lakh rupees per store per month. A 2.5x gap on the per-store revenue, propagating to an 8x gap on the affordable marketing envelope.

The corrected Q2 cash investment landed at 5.93 to 8.23 lakh rupees. Pack 1 Meta at 3 to 4.5 lakh rupees projected at 1.3 to 1.4x direct ROAS. Pack 3 Google at 2.25 to 3 lakh rupees projected at 1.4 to 1.5x. Both packs landed right at the EBITDA-positive threshold for Indian QSR variable cost. The original envelope, had it shipped, would have torched roughly 45 lakh rupees of cash against a margin structure that could not absorb it. The single corrective action was pulling per-store revenue from the database before writing the spend plan.

Why the right number rarely exists in PowerPoint

Three habits explain how founder-stated numbers drift from operational reality.

Habit one: the count is wrong. The same F&B brand had internally reported 199 stores until a DB pull surfaced 86 active stores (51 COCO and 35 FOFO), the remainder being closed FOFOs inflating the network count. A marketing plan sized against 199 stores is sized against capacity that does not exist. The fix is to anchor on operational data, not deck data.

Habit two: the seasonal mix is averaged out. A per-store monthly average rolls together festive weeks and dead weeks. A Q2 plan needs the Q2-specific number, not the trailing 12-month mean. Pull the same months of the prior year, adjust for new stores, then size.

Habit three: the product mix is invisible at the headline level. The same DB pull surfaced the unit economics that matter: a Choco Heaven Waffle product moving 9,741 units at 134 rupees AOV, and a Holland Pancakes category at 1.096 crore rupees revenue across 58,404 units. Marketing spend on a discontinued or thin-margin product is wasted regardless of how clever the creative is. The product mix sets the upper bound on what any campaign can return.

The four numbers that have to be pulled before the envelope is written

Evidence from a BFSI engagement where the number was missing

The same discipline applies to BFSI lead-gen. A $6 million instant-loan fintech was running 100 percent on paid acquisition, 28 paid keywords driving over 1.1 million monthly impressions, and 526 ranking keywords organically of which roughly 470 were branded. The marketing question the brand was asking was “how do we scale paid faster.” The unit-economics question, which had not been asked, was: what is the cost per funded loan after rejection, after fraud, after early default, by channel? Without that number, scaling paid simply scaled the unprofitable customers along with the profitable ones. The audit recommendation was to instrument funded-loan attribution by channel before raising the paid budget another rupee. See the paid search audit for how this gets folded in.

Why this matters even more in 2026

Three forces have raised the cost of getting unit economics wrong. First, paid search auction prices on the head terms in fintech, real-estate, and education have climbed faster than category revenue growth over the past four quarters. The brands paying premium CPCs on a thin margin per acquired customer are bleeding. Second, AI Overviews are absorbing informational queries that previously fed top-of-funnel organic traffic, so brands cannot recover bad paid economics through organic blog volume the way they could three years ago. Third, attribution is harder, not easier. LLM-mediated discovery now produces an “unattributed branded search” two to four weeks after the model citation that doesn’t map to any campaign. See attribution modeling when LLM traffic is untrackable for that thread. All three pressures favour the brands who know their unit economics cold.

Practitioner takeaway: five actions before the next budget review

1. Pull per-unit revenue from the source system. POS, CRM, billing, or accounting. Not from a slide. Not from a stated baseline.
2. Map net contribution margin per SKU or per customer cohort. The marketing budget cannot exceed what the contribution margin can absorb.
3. Calculate payback per channel. Paid social, paid search, organic search, content. If any channel’s payback is longer than customer life, throttle it.
4. Anchor the spend envelope on the bottom-up number. Then test it against the top-down rule of thumb. If the two disagree by more than 30 percent, the top-down number is wrong.
5. Re-run quarterly. Unit economics drift as product mix, channel mix, and competitive auction prices change. A number that was right in Q1 may not be right in Q3.

FAQ

What if the brand does not have clean operational data?

Most multi-location and multi-SKU brands do not, at the start. The 86-store F&B brand needed a Laravel command centre with Rista POS sync, a 12-month historical backfill, and supervisord-managed workers (after a 3.5 day silent worker death) before the unit economics were reliable. Cleaning the data is the prerequisite to using it. If the operational data is unavailable, the budget envelope is conjecture.

How does this differ for SaaS versus retail?

The numbers shift. SaaS uses MRR per customer, gross margin (typically 75 to 85 percent), CAC payback period, and net revenue retention. Retail uses revenue per store, contribution margin per SKU, basket size, frequency, and per-store fixed cost. The discipline is the same. Pull the numbers from the operating system before sizing the spend.

Is “percentage of revenue” ever the right anchor?

It is a useful sanity check, not a primary input. A 10 percent of revenue marketing budget on a 5 percent net margin business is materially different from the same percentage on a 30 percent net margin business. The contribution margin governs what marketing can afford. The percentage of revenue is a coarser proxy that hides that.

How often is the founder-stated number wrong by more than 2x?

On the F&B engagement: 2.5x off on per-store revenue, 8x off on the affordable envelope. Across the multi-location, multi-product engagements ScaleGrowth has run, the founder-stated number sits within 25 percent of operational reality in roughly one in three cases. The other two in three justify the database pull on their own.

Get the diagnostic

If your next marketing budget review is more than three weeks away, that is enough time to pull the per-unit revenue, contribution margin, channel payback, and quarterly seasonality from the operating system. Request a growth strategy consultation, and the unit-economics map gets built before the spend envelope is drafted.