ROI Is Dead: Now Bury It

Advertising and advertisers are being increasingly held to account. Quite right too. 100 years ago CEOs started worrying about which half of their advertising was wasted, and it is high time we came to terms with the financial analysis of advertising, and indeed marketing, performance. In particular we need to embrace good practice and discard the bad. We need to bury the bad, notably ROI. I will get to that, but first a few good practices.

Discarding sales volume as the yardstick of advertising performance is progress. If we can, we should identify changes in consumer behaviour resulting from the ads, then use profits rather than sales, because advertising typically influences price more than volume. Sales volume incurs variable costs, but price drops almost directly to the bottom line. Advertisers should be rewarded for price effects. Better still, as the IPA has pointed out [1], use shareholder value as the performance indicator, since brand advertising affects investors and analysts as well as customers.

If changes in customer behaviour are not directly attributable to advertising, we can look to changes in brand equity as measured by intermediate effects such as intention to buy, or relevance, or perceived quality, or saliency. If brand equity does not show positive changes either, then we can still justify the cost of advertising as maintenance. We do not paint the warehouse because it has a positive ROI, but because not painting it will cost us more in due course. And who calculates the ROI of the finance department?

Agree the business model

Financially motivated CEOs are increasingly reluctant to accept these intermediate effects and arguments in isolation. Since the CEO is both correct and the boss, staking out the two 'show me' and 'give me the money or we die' positions in opposition solves nothing. The solution lies in formulating andagreeing the firm's business model. Every organisation needs to source and harvest cash flow. That, in essence, is marketing. In some business sectors, such as B2B, advertising plays little part, but every business needs to build an asset called goodwill, or reputation, or brand equity, which in turn facilitates sales volume and the price customers are prepared to pay. These in turn drive the bottom line. Marketing itself is not an option, but the balance of expenditure, immediate profits and brand equity (later profits) is.

The key is to get away from bargaining over the size of the advertising budget and get into a joint study of the business model. Using discounted cash flows (DCF) to evaluate the options is good business practice even though the variables cannot be known. The team can look at various estimates and home in on the believable, feasible and affordable.

Somewhere along the track the CEO typically asks what success will look like. The question takes us from deciding the plan to assessing performance. These are two very different processes: the first is about the future and the second is about the past to date. Most companies now recognise that assessing advertising performance to date needs both financial and non-financial measures. These are not 'hard' and 'soft'. Numbers do not become hard just by putting $ signs in front of them. Hard numbers are the ones that have been professionally gathered and collated, and are reliable.

Curiously enough, two of the most popular metrics are also the softest, ie unreliable, and both are financial: ROI and DCF, when applied to performance assessment. Let us take them in turn.

Why ROI should not be used as an advertising metric

ROI was devised for assessing capital projects where the investment is made once and the returns flow during the following years. Marketing people like to say that advertising is an investment, in order not to be treated like any other cost. In a way this is right: advertising is an investment in the future cash flow; and the benefits often extend beyond the current year. At the same time, it is not an investment in the capital project sense. Politicians refer to any expenditure they favour as an investment. Do marketers want to be tarred with the same brush?

Advertising expenditure is usually continuous from year to year and, mostly, maintains the brand and the bottom line. It belongs in the profit and loss account, not the balance sheet. ROI does not cope well with ongoing future budgets: it is the short-term return divided by the short-term advertising expenditure. If advertising is a single, one-off, activity and the resultant future flows of income can be DCFed to present value, then a slightly better case for ROI can be made but, as we will see, it is still wrong.

The profit, or economic value added, or increase in shareholder value from advertising all require the costs to be deducted from sales revenue, along with the other costs. ROI, however, is the net profit return (R) divided by the advertising investment (I). This arithmetic difference between subtraction and division lies at the heart of the problems with using ROI as a performance metric.

Making this point to marketers all too often meets the response 'subtraction or division – what difference does it make?', which says more about the numeracy of some marketers than I care to think about.

The example shows that ROI gives a false picture, partly as a result of excluding longer-term cash flows, the effects on the dynamism of the business and brand equity. In fact, a drive for increased ROI will ultimately destroy any business, because the cash-generating activity is penalised relative to short-term profit. The example is typical: reducing advertising expenditure usually increases ROI.

Another typical response in support of using ROI is that the ratio is useful for comparing alternative uses of the same budget. In fact, if the I is constant, then profit peaks at the same point that ROI does, so the ratio is still redundant at best and possibly misleading. The immediate reaction to a high ratio is the expectation that more should be spent – which means that the I is no longer constant and the argument fails.

In other words, stretching to improve ROI is sub-optimal in terms of the firm's goals (profitability, cash flow or shareholder value) and will typically cause marketing expenditure to be lower than the firm should want it to be. The reality is that investors, analysts and other stakeholders depend on profit (R minus I) not on an artificial ratio of the two numbers. Payback, profits and shareholder value are all reputable metrics and all subtract costs from revenue. Professional accountants should know all about this, since it is taught in basic accounting classes. They typically prefer DCF calculations such as the present values of alternative expenditure strategies, or shareholder value. This is good  practice in planning but not for evaluating performance to date, for a different kind of reason.

Discounted cash flow

DCF is about the future, not the past. It requires the evaluator to estimate customer and competitor behaviour for years ahead. The evaluator has to estimate the strength of the economy, customer demand and category performance in those years, as well as how productive the advertising will be. Since planning is also about the future, the technique is consistent, albeit highly conjectural. To use it to estimate the present position of the brand, however, is inconsistent and illogical. Suppose you are sitting in a plane and the pilot comes on the intercom to say he does not know where the plane is now but he has estimates of where it will be and by back-tracking from there he has a pretty good idea of where you might be. Would you worry about his competence? Locating the present position of a plane, or of a brand, requires present indicators or, failing that, dead reckoning of the path already taken.

A number of techniques have recently become fashionable in academic circles based around DCF estimates, such as brand valuation[2], Customer Lifetime Value and 'customer equity' [3]. These can be useful, as noted above, for planning purposes, but not for performance assessment.

If ROI is dead, why is it not buried?

It is odd that ROI has taken such a prominent role in the recent calls for accountability. Advertisers should indeed be able to justify the expenditure in financial terms, ideally the impact on the bottom line. One can call that productivity or payback that is the return from (minus) expenditure, but the ROI ratio is misleading.

Colin Farrington, director general of the Institute of Public Relations, suggested that the term is a fashion statement. PR people use it generically without knowing what it means: 'Ask ten PRs to define ROI and you'll get ten different answers.' Without getting into the detail above, he claimed, 'the obsession with return on investment is damaging PR's effort to demonstrate the value of its work'. He is right; and the same can be said of advertising or any other part of the marketing mix.

Another example is from page 31 of the July/August 2004 Admap, where the author states that 'any positive change in NPV for increased spending represents a positive ROI.' If, however, one inspects his data, it appears that the greatest increase in Net Present Value gives an ROI of 120%, whereas a 20% lower increase in NPV gives an ROI of 167%. In other words, as net profit increases, ROI goes down. The author is presumably employing the term ROI as if it was payback.

One more example comes from the cover of the July/August issue of Admapin 2003: 'How to budget for a healthier ROI.' Study of the article reveals that ROI is not one of the 'key measures' or 'strategic value measurements' at all. The section on 'Strategic ROI' calls for it to be calculated over the long term 'using financial and non-financial measures, focusing not just on the short term today but also on where long-term brand health will be generated'. That is a fine objective, but you cannot calculate ROI that way.

Warding off evil spirits

ROI is not so much understood as waved about as a totem to ward off evil spirits, namely those trying to cut advertising expenditure: 'We are using ROI to establish the budget, so leave us alone.' Look behind these claims and you are unlikely to find much substance. Direct marketing is an exception. If samples are matched and only one of each receives the direct marketing treatment, then an ROI calculation may be feasible, although other factors have a way of creeping in. In theory, the same can be done for mass advertising, but in practice it is very difficult. Much more often the ROI claim is bogus.

The totemic use of ROI may work for a time, just as placebos work in medicine, but using a false metric in this way will bring as much discredit on marketers as the sellers of bogus medicines brought on themselves in the 19th century. ROI is dead and now it should be decently buried. If advertisers seek respect for their proposals, then they should use respectable metrics.

Example: The Brown Bread company

The Brown Bread company is trading comfortably and is looking for modest growth. As the new finance director is asking the marketing people to maximise ROI, they produced the three plans in the table below. The net profit is given by the sales revenue after deducting the marketing spend, and variable and fixed costs. As far as they can judge, and the numbers are not challenged by their colleagues, the first maximises profit and cash flow, and the second significantly enhances ROI. A third option maximises ROI in the sense that, with no marketing expenditure at all, ROI is infinite. Which should they choose? (See Table 1.)

Plan B puts the sales into decline after some years of increase with the probability of a lower base in future years. Profit declines significantly – but ROI increases by 5%. Greater cuts of marketing expenditure (Plan C) increase short-term profits considerably and also increase ROI – at the expense of sales revenue.