Return on Marketing Investment Explained
Anybody can produce a flattering return figure without stating a single untruth. The formula comes first here, then the rest of the page explains where it misleads and what to ask about any number somebody puts in front of you.
Return On Marketing Investment Defined
Return on marketing investment is the profit generated by marketing, less the cost of that marketing, divided by the cost of that marketing. It answers one question: for every pound spent, how much came back beyond the pound.
What sits in each part. The detail decides everything.
The profit generated, not the revenue. The cost of the marketing, meaning all of it rather than only the media. Almost every dispute about a return figure is a dispute about one of those two.
Return on investment. The general business measure.
Applied to any expenditure. Marketing return is a specific application of it. The two get used interchangeably by people who mean different things.
Return on advertising spend. Narrower, much more common.
Revenue divided by advertising spend. It excludes the cost of producing the advertising, excludes fees, excludes internal time and uses revenue rather than profit.
Which one you are usually shown. The third.
It is the easiest to calculate, it is what most advertising platforms report by default and it produces the largest number of the three. None of that makes it dishonest. It does make it the least informative.
A Worked Example
One campaign, calculated properly. Every figure is invented for illustration and every input is an assumption. Replace them with your own.
The costs. All of them.
Assumption: £4,000 of media spend. Assumption: £1,000 in fees. Assumption: £500 of production, being photography and page building. Assumption: £500 of internal time. Total cost £6,000.
The result. Converted to profit.
Assumption: the campaign produced £30,000 of revenue. Assumption: the gross margin is thirty percent, so those sales contributed £9,000 of gross profit.
The calculation. Profit less cost, over cost.
£9,000 less £6,000 is £3,000. Divided by the £6,000 spent, the return is 0.5, being fifty pence of profit for every pound invested.
Why gross profit is the numerator. Revenue is not yours.
Most of that £30,000 pays for delivering the work. Using revenue would report this campaign as returning four pounds for every pound spent, which describes a business that could be losing money on every sale.
What the figure actually says. It worked, modestly.
A positive return calculated this way is a real result. It looks far less impressive than the same campaign reported the usual way.
Why Revenue Based Figures Flatter
Take the campaign above and report it the way it usually gets reported. Nothing changes except which numbers are used. The answer becomes unrecognisable.
The usual method. Revenue over media spend.
£30,000 of revenue divided by £4,000 of media spend gives seven and a half to one. That is a figure anybody would be pleased to present, drawn from exactly the same campaign that returned fifty pence in the pound.
What produced the gap. Two omissions.
Revenue rather than profit, which ignores the cost of delivering the work. Media spend rather than total cost, which ignores fees, production and internal time.
Why this matters at thin margins. The sign can flip.
A business at a slim margin can report an impressive revenue based return on a campaign that lost money once the cost of delivery is counted. The reporting is arithmetically correct throughout.
The question to ask. Two sentences.
What is in the top of that fraction, then what is in the bottom. A supplier who answers immediately is using a defined method. One who has to check is reporting whatever the platform showed them.
Attribution, And Why The Same Campaign Has Three Answers
When somebody sees an advert, then searches, then returns by email and buys, which of those gets the sale? Attribution is the rule that decides. The rule is a choice.
Last click. All credit to the final step.
The most common default. Over credits whatever closes, which is usually search and email, while crediting nothing to what created the interest.
First click. All credit to the introduction.
Over credits awareness channels and ignores everything that turned interest into a decision.
Linear. Split evenly across every step.
Fair in the sense that nothing is ignored. It also treats a passing impression as equal to the conversation that closed the sale.
Position based. Weighted to the first and last.
A compromise that credits introduction and closure most, on the reasonable assumption that those matter more.
Data driven. The model infers weightings from observed paths.
Better in principle and harder to interrogate, because you cannot easily see why it decided what it decided.
The rule that matters. Choose before the results arrive.
A model selected afterwards is a model chosen because of the answer it gives. Fix it in advance, then keep it, since changing model between periods makes every comparison meaningless.
What Is Hard To Measure
Some of what marketing produces cannot be attributed at all. Pretending otherwise is how reporting becomes fiction, so each of these gets a practical handling rather than a solution.
Telephone enquiries. Trackable, with effort.
Call tracking connects a call to a source. Without it, a business taking most enquiries by phone is reporting on a minority of itself.
Walk ins. Ask them.
No system captures this. One question at the counter, recorded consistently, produces rough data that beats none.
Offline closes. The gap between enquiry and sale.
Marketing reports enquiries, the business closes them elsewhere, then nobody joins the two. Recording the source against the won job is the fix. It is a sales discipline rather than a marketing one.
Long sales cycles. Outrunning the window.
If the decision takes longer than the reporting window, the sale falls outside the period that paid for it. Widen the window to match the cycle.
Word of mouth started by an advert. Unmeasurable.
Accept it as unrecorded upside rather than trying to estimate it, since any estimate becomes a number somebody later treats as fact.
Lifetime Value
For a business where customers return, judging a campaign on the first purchase understates it substantially. Bringing repeat value in is correct. It is also where optimistic assumptions enter the calculation.
Why first order return misleads. The first sale is not the relationship.
A customer who returns three times is worth three sales for the cost of acquiring one. Judged on the first transaction, the campaign that won them can look like a failure.
Where it applies. Repeat purchase, subscriptions, retainers, servicing.
Where it does not. Genuinely one off purchases.
A business selling something people buy once in a decade cannot use this. Applying it anyway is where the abuse starts.
How to do it without inventing. Use observed retention.
What your customers have actually done, from your own records. Not what you hope they will do, nor what the sector supposedly does.
The danger to name. Optimistic retention justifies anything.
Assume customers stay a little longer than they do and any acquisition cost becomes defensible. That is how businesses spend heavily against revenue that never arrives.
Payback Period
How long before the money comes back. For an owner this is often the more important question, because it is about cash rather than about profitability.
What it measures. Time, not size.
How long the money is out of the business before the customer has repaid what it cost to win them.
Why it can override a return figure. Cash is a constraint.
A campaign returning well over two years may be unusable for a business that needs the money back this quarter. A weaker campaign paying back in weeks may be the correct choice. No return ratio will tell you that.
How to calculate it. Cost against contribution over time.
Take the acquisition cost per customer, then work out how many months of that customer's contribution are needed to cover it.
When to lead with it. Tight cash, growth phases, anything financed.
Any business where the timing of money matters as much as the amount. Which, for most small businesses, is most of the time. That is why this belongs alongside the budget rather than after it.
Would It Have Happened Anyway
Every return figure assumes the marketing caused the sale. Incrementality asks how many of those sales would have occurred without it. It is the question that separates real measurement from bookkeeping.
What incremental means. Additional, rather than recorded.
A campaign can be credited with a hundred sales and have caused twenty. The other eighty were coming. The advertising happened to be present on the way.
Where this bites hardest. Brand terms and remarketing.
Advertising against your own name reaches people already looking for you. Remarketing reaches people who already visited. Both report excellent returns for the same reason: they target people who were already coming.
The uncomfortable implication. The best looking line may be the least valuable.
Which does not mean stopping either. It means knowing that a portion of that reported return is bookkeeping rather than growth.
How to test it. Hold something out.
Pause the activity in one region while keeping it elsewhere, else switch it off for a defined period, then compare. Crude, disruptive and the only way to find out.
What this is worth. A whole budget line, sometimes.
Businesses that run this test frequently discover a meaningful share of their spend is buying sales they already had.
Setting A Target
A return target is a commercial decision rather than an industry standard. Three things set it. None is what other businesses achieve.
Margin sets the floor. Break even, first.
Work out the point at which a campaign covers its own cost including delivery. Anything below that is buying revenue at a loss, whether or not it looks busy.
Growth appetite sets the ambition. Above break even, deliberately.
A business that wants to grow accepts a lower return to buy volume. That is a legitimate decision, provided it was decided rather than discovered.
Cash position sets the ceiling. What you can survive.
How long you can fund the gap between spending and being paid. This constrains the plan regardless of how good the return looks on paper.
Buying growth on purpose. Named as such.
A target below break even is a decision to buy market position with money. Perfectly reasonable, provided it is written down as that rather than reported as a return that happens to be poor.
Reporting It Properly
An owner or a board should see enquiries, cost per enquiry, sales, cost per sale and the return calculated on profit. Five questions interrogate any report. They apply to agencies including ours.
What counts as a conversion. The first question, always.
An enquiry sent, else somebody reaching a page. Those produce very different numbers under the same heading.
What attribution window is set. How long after contact a sale still counts.
A long window flatters. A window shorter than your sales cycle understates. Either can be defended. Either should be stated rather than discovered.
Is view through included. Sales credited without a click.
Where somebody was served an advert and bought later without ever clicking. Ask whether it is in the figure, because it changes the number substantially.
Are brand terms separated. The one that reveals most.
Advertising against your own name should be reported separately, since mixing it in raises the average while telling you nothing about reaching anybody new.
Are fees inside the cost figure. Including the reporting agency's own.
A return calculated on media spend alone excludes the cost of the people producing the report. That is the most common omission in the industry. It is worth asking about directly.
Common Mistakes
Each of these produces a number that is arithmetically correct and commercially misleading.
Counting revenue rather than margin. The single commonest error, able to turn a loss into an apparent success at thin margins.
Excluding fees and internal time. Counting only what went to platforms. The work of running the campaign was real and somebody paid for it.
Mixing attribution models between periods. Comparing a quarter measured one way against a quarter measured another. The change in the number is the method rather than the performance.
Judging on a window shorter than the sales cycle. Reporting monthly on a decision that takes a quarter. Every campaign looks like a failure at the point it is assessed, so the good ones get stopped, per the marketing plan.
Treating a ratio as a result. The one that matters most. A return figure is a question about what was included, which attribution was used and what would have happened anyway. Display reporting in particular rewards this scrutiny, per what is display advertising. The measurement itself should be checked via an audit. The full series is on the digital marketing guide.
