A Tale of Maths and Ignorance: Why ROI Is Not Effectiveness


I’d like to see the ROI on that spot”
(Overheard in the blogosphere)

There’s a peculiar tendency – particularly amongst agency folk – when people want to appear serious, grown up, responsible, and accountable. They talk about the ‘ROI’ – return on investment.

While they might think ‘ROI’ is just another phrase that broadly indicates effectiveness or payback, unfortunately the mere utterance of phrase ‘ROI’ often identifies that person as someone who doesn’t know what they’re talking about.

Simply put, ROI is a ratio – it represents the short-term return divided by the short-term advertising expenditure. For example, if £3 profit were generated for every £1 of marketing expenditure the ROI would be expressed as 3:1.

ROI requires the profit to be divided by the expenditure. It is the net profit return (R) divided by the advertising investment (I). And the fact that ROI is derived through division is important. It is a measure of efficiency. ROI in other words is a measure of how hard working your marketing investment is. How much bang for each buck you get, if you like.

But (and it’s a big one) it is not a measure of effectiveness.

Any calculation of effectiveness – that is the economic value created – requires all the associated costs (media, production, etc.) to be subtracted from sales revenue. In other words calculating effectiveness means we should be calculating R minus I, not merely R divided by I.

This is something that seems to be lost on a great many of those who bandy the term ‘ROI’ about so freely. Clients and agencies alike. Certainly if strategists of whatever kind are failing to understand the distinction between effectiveness and efficiency (and by implication between division and subtraction) one might be forgiven for questioning their grasp of all things strategic. And whether they are deserving of a place on the payroll.

This distinction between subtraction and division is not pedantry. It matters, because blindly chasing ROI can actually destroy economic value for a brand or business.

Compare and contrast these two scenarios, for example:

Scenario 1: A brand spends £2m and generate £6m sales

Scenario 2: A brand spends £5m and generates £10m sales

ROI chasers would of course go for the Scenario 1. It generates £3 for every £1 spent. An ROI of 3:1.

However, Scenario 2 actually delivers more sales and more profit. Even though its ROI was lower, generating £2 for every £1 spent.

So, which company do you want to be a shareholder in? Or an employee of? Or submitting an effectiveness awards paper for?

There is one other significant problem with ROI. It represents the short-term return divided by the short-term advertising expenditure. The problematic word in all of of this is of course, ‘short’.

Charles Channon had long ago warned of the lure of efficiency: “It works in the short term ( the bottom line or the decision maker do not have to wait for the benefit), it is measurement-friendly (and some would say measurement-led) and it is highly actionable (in that it appears to offer a greater degree of certainty in both planning what to do and evaluating it afterwards).”

But as Peter Field has put it: “The single greatest threat facing marketing at the moment is short-termism.”

ROI does not take into account the longer and broader effects of advertising. It takes not take into account longer-term cash flows, and the effects on the dynamism of the business and brand equity.

The evidence is that most advertising activity does have some kind of short-term volume effect. So far, so good. However, the average uplift is small – usually about 5% in grocery markets. That is too small to be profitable for all but the biggest of brands.

Short-term effects, whether from advertising or price promotion, are usually unprofitable. The economic value that advertising creates emerges from long-term effects: gaining new customers who go on to buy again and again at premium prices.

Tim Boadbent cites a study of single-source data that measured advertising’s effects over 12 months as 2.5 times larger than the initial sales uplift. Over a full year, repeat buying means that the average 5% uplift in short-term sales accounts for about 12% of the brand’s total annual sales.

Broadbent reminds us that because a typical grocery brand accounts for about only 40% of a household’s category buying, attracting new users has a significant impact on long-term sales. IRI for example, have measured the carry-over effects in years two and three as almost equal to the effects in year one, namely 2.4 times the uplift. Thus, the short-term 5% sales uplift is multiplied to about 29% of the brands volume by year three.

Clearly, we should be calculating the long-term effects of advertising, not merely the short-term effects.

So if you want to assess the short term efficiency of your efforts, by all means measure ROI. Calculate R divided by I. Conscious and clear sighted as to its limitations.

And if you want to calculate the financial value your efforts create over the long term, measure the effectiveness of your efforts. And calculate R minus I.

But please, don’t speak of them as if they’re interchangeable.


Tim Broadbent, ‘How Advertising Pays Back’, Admap, November 2001, Issue 422

Charles Channon, ‘The differences between effectiveness and efficiency’, Admap, March 1990

Peter Field, The Warc Blog:  ‘Short-termism: The single greatest threat to marketing’

John Philip Jones: When Ads Work, 1995

Colin McDonald: ‘How frequently should you advertise?’ Admap, July/August 1996

One comment

  1. Bill Carlson

    As “agency folk” with a strong understanding of financials, I don’t get hung up on the specifics of the term but rather the intention since the term is nothing more than shorthand for discussions of payback and accountability — those that take it overly literally and pull out their college accounting textbooks to talk about it can own their own perceptions but I would argue they’re over-thinking it.
    That being said, I take issue with the tone here and can say with certainty that it’s by no means a “peculiar tendency” of agencies to want to provide a prediction/forecast of success to justify certain marketing investments — that is pretty clearly a client-side demand which responsible agencies will certainly strive to address.
    And those questions don’t come only from the CFO — every senior manager needs to justify expenditures to someone and they have only so much to work with and therefore must rationalize the priortization of spending. And (smart) agencies recognize the need to look at the excitement of brand strategy and creative expression through a financial lens — at the end of the day, an agency that delivers great work with no positive changes in their client’s business will find themselves on the short end of the “agency review” stick. (And yes, some debate there as to whether an agency’s work is actually “great” if it doesn’t impact results, but that’s a different discussion.)
    Interestingly, one of the challenges for “agency folk” is a client who wants to have some sense of projection of results but who won’t finance the necessary data-gathering (e.g. often some sort of research but could be other things like supporting an A-B geography or store test or internal tracking of pre vs. post sales, etc., to establish a measurable understanding of before and after). Whether it’s a cost or time consideration, often a combo of both, the best that can often be done is some “well-intentioned guesstimating” — a precarious position for the agency as they rarely have control over all the dynamics of the situation…
    And finally, your two scenarios fail to take into consideration time vs. limited availablity of capital. So 2 may get you 6 and 5 may get you 10, and higher volume is a good thing in many non-marketing ways such as improving the allocation of fixed-cost overhead, but if all there is to spend is 2, it’s good to know how to spend that 2 effectively.