How do you measure marketing ROI as a business owner?

The marketing ROI calculation that works for owners

The owner-grade formula is blended return, often called MER (marketing efficiency ratio): total revenue divided by total marketing spend — all channels, all fees, one number. If you spent 10,000 on ads and agency fees this month and generated 42,000 in revenue, your blended return is 4.2. No attribution model, no platform's opinion, nothing to game; it moves only when reality moves.

Two refinements make it decision-grade. First, run it on gross profit, not revenue, when your margins vary — a 4.2 on 60% margin products and a 4.2 on 15% margin products are entirely different businesses. Second, know your break-even blend: divide 1 by your gross margin. At a 40% margin, you break even at 2.5 — anything above that earns, anything below pays the platforms for the privilege of working. That one number turns every future marketing conversation into a yes/no question.

The four marketing metrics for owners — and the vanity ones to drop

Four numbers, tracked weekly, cover what an owner actually controls:

• Total marketing spend — ads plus fees plus tools, the real all-in number.
• New-customer revenue — kept separate from repeat revenue, so acquisition is judged on what it actually acquired.
• Cost per acquired customer (CAC) — spend divided by new customers; compare it to what a customer is worth over time, not just their first order.
• Blended return (MER) — the referee from the section above, ideally margin-adjusted.
• Drop from the owner dashboard: impressions, reach, likes, follower counts, and each platform's self-reported ROAS on its own. They are diagnostics for whoever manages the campaigns — useful one level down, misleading at the decision level.
• One caveat that saves arguments: platform ROAS is still valid for comparisons inside a platform — campaign A versus campaign B under the same measurement bias. It fails only when used as absolute truth or compared across platforms.

Attribution for a small business: why no model is perfect

Attribution tries to answer 'which touchpoint deserves credit?' — and every model is a defensible opinion, not a fact. A customer sees your Snapchat ad, googles you a week later, clicks the search ad, then messages on WhatsApp: Snapchat, Google and your WhatsApp agent will each claim the sale, and the platforms' combined claimed conversions will exceed your actual sales. That is structural, not fraud.

A small business should not buy its way out of this with attribution software; it should triangulate three cheap signals: the blended math (which cannot lie), a 'how did you hear about us?' question asked at every enquiry and logged honestly, and UTM-tagged links or unique WhatsApp numbers per channel where practical. When all three point the same direction, act. When they disagree, trust the blend first, the customers' answers second, and the platforms last.

The weekly dashboard: answering 'is my marketing working?' in 15 minutes

One spreadsheet tab, one row per week, filled every Monday from the same sources — ad managers, bank or POS, and your enquiry log. Columns: total spend, enquiries or orders, new customers, new-customer revenue, CAC, blended return, and one line of notes for anything unusual (an offer, a holiday, a stockout, a tracking change). Fifteen minutes, no design, no dashboard software required.

The value is in the trend, not the row. Week-to-week noise is normal — Gulf demand swings hard around Ramadan, Eid, summer travel and salary days — so judge movement over four-week windows, and write the note the week something odd happens, not a month later when nobody remembers. After a quarter, this one tab answers 'is my marketing working?' faster and more honestly than any agency report, because it is denominated in money that reached you.

When do you kill a marketing channel? Decision rules that remove the emotion

Channels die badly in two ways: too early, murdered in week two before the platform finished learning, or too late, kept alive for a year because someone liked the idea of 'being on TikTok'. Written rules prevent both. Give every new channel a fair test window — four to six weeks and enough budget for the platform to see real conversion volume — and define the pass threshold before launch, anchored to your break-even blend.

Then apply a two-strike rule: if the channel misses its threshold for two consecutive review windows and nobody can name a specific, testable fix — a new creative angle, a different offer, a landing page change — it dies, and its budget moves to the best performer the same week. Two honest exceptions: don't execute a channel during an obvious seasonal trough it didn't cause, and don't judge anything in the same window when tracking was broken. Everything else is sentiment, and sentiment is the most expensive line item in most ad accounts.