Is There a Silver Metric for Marketing Accountability? (Pt. 1)

By Tim Ambler, London Business School 

Performance, be it business overall or marketing specifically, is essentially multidimensional: Superior performance against one objective cannot easily be traded off against lack of performance on another. Short- and long-term profits cannot be satisfactorily merged into a single number because we need short-term cash flow, however attractive the long term may be, and because the short term, once successfully achieved, does not guarantee the long term.

Even so, some executives see marketing as an investment like any other that should be evaluated by the return on that investment (ROI) or some variant on the incremental discounted cash flow (DCF) that can be expressed in various ways, e.g., increased net present value (NPV), customer equity, brand valuation, shareholder value, or customer lifetime value.* ROI aside, these are all different labels for the same thing, namely incremental cash flow in future years, discounted by the time costs of waiting for the money and added together. Risk factors are sometimes included. For simplicity, we will refer to this group as "NPV."

The American Marketing Association defines marketing accountability as: "The responsibility for the systematic management of marketing resources and processes to achieve measurable gains in return on marketing investment and increased marketing efficiency, while maintaining quality and increasing the value of the corporation."

The significant part of that definition is the duality of short-term gains and enhancing the quality and value of the corporation, i.e., brand equity or the intangible asset created by good marketing. Properly evaluating marketing performance seems, therefore, to require both to be considered.

Any single metric in isolation is almost meaningless. At a minimum, we might need to know:

In a way, brand health is like human health. If blood pressure is normal but cholesterol levels are exceptionally high, a single index averaging the two would likely understate the potential calamity.

Can We Use NPV as a Performance Metric?
When considering alternative future strategies and marketing expenditures, NPV comparisons make sense, at least as part of the consideration. Comparing NPV estimates made in one year with those made in other years is complicated by the different expectations of interest rates (the time value of money), market growth, or by economic and risk indicators at the dates when the different estimates were made. When planning, however, all NPV outcomes are estimated at the same time, which enables these contextual factors to be standardized. So NPV may be a crucial part of marketing planning, but that is not performance evaluation. One looks forward and the other considers the past to date.

Estimating future cash flows is a logical way to evaluate future activities, but looking at the future to evaluate the past is another matter. Looking at tea leaves may be more useful. Assessing past performance involves comparing what has taken place and where we are now with what should have taken place and where we should be.

 

Yet the problems with using future NPV, in whatever form, for evaluating past performance go deeper than that. Suppose we have the perfect vision of future cash flows. Further suppose that the CMO has performed so badly to date that he is replaced at the beginning of the first future period by a brilliant new performer. If the poor cash flow for the last period is aggregated with the now excellent future performance, the overall NPV will exceed that of a merely average CMO. In other words, we have just persuaded ourselves that the bad performance we have seen is really rather good. Bring back the sacked CMO! In this logic, his poor performance not only created great future performance, but, because NPV is being used to assess past performance, that looks great too.

Thus using NPV for to-date evaluation means taking credit for future performance not yet achieved. And there are infinite possible futures. Which one would you like to pick? How will future changes in the customer base be assessed? Which customers will be acquired and lost with what impacts on cash flow? And how many new customer acquisitions will be due to past actions, future actions, or both? In businesses with high churn rates, such variation would be especially difficult to assess, not least because future churn depends on future marketing actions. Yet the NPV approach requires us to use those estimates to assess past performance.

Another thought experiment concerns a yacht race. If two identical yachts leave the same London port at the same time, both headed for New York, the yacht closest to the destination must have performed better to date, right? Here we are using our expectation of future performance, namely getting to New York first, to assess performance to date. Yet the example is illusory. The estimate of the remaining race time is a forward-looking view based on present locations and how long each will take, from their current relative coordinates, with assumptions being made about winds, currents, and tides. Performance to date would be better assessed by comparing the advantages of their current locations vs. the expectation or plan.

 

In summary, NPV metrics are valuable for planning purposes but not for evaluating performance to date. The reason is simple: We can forecast the future in many ways but we cannot measure it.

Return on Investment
Return on investment (ROI) was devised for comparing capital projects in which an investment is made once and the returns flow during the following years or periods. ROI is the net return divided by the investment, or, more correctly, the incremental profit as a ratio of the incremental expenditure. Although in theory ROI can be calculated using NPVs, in practice estimating the differential cash flows beyond the present year is difficult and ROI is usually calculated using the annual profit and expenditure either in whole or incremental portions.

The first problem with ROI as a marketing metric is whether marketing expenditure is an "investment" in the original sense. For capital projects, cash flows in for some years and the initial one-time investment will typically result in some residual value. ROI for a proposed investment is expected to exceed alternative potential uses for the money, such as a certificate of deposit or the buy-down of existing debt, but not to pay back in a year or two. Marketing expenditure, however, has more to do with maintenance of the asset, brand equity, than creating a new asset (investment).

 

The second problem is that ROI requires the profit to be divided by the expenditure, whereas all other bottom-line performance measures consider profit or cash flow after subtracting expenditure. Division rather than subtraction distorts the metric, as higher ROI can be achieved just by spending less, not necessarily achieving more. The profit or economic value added from marketing requires that costs be subtracted from sales revenue.

The third problem is that pursuing higher ROI as a target can cause underperformance and sub-optimal levels of activity. This arises from the law of diminishing returns. After the point of the profit response curve where ROI is maximized, further sales will typically still make a profit, albeit at a diminishing rate. There are a few rare examples of ROI and profit both maximizing at the same point. One might be a seller of ice creams on a beach where 30% of those present are on no-ice-cream diets, but the remaining 70% are happy to buy one item each but no more. If the seller's marketing costs are low, then both profit and ROI are maximized at 70% penetration since a ceiling has been reached.

The fourth problem with ROI is that the baseline, i.e., what would have happened without the expenditure, is hard to determine and mostly likely to be subjective judgment and vulnerable to manipulation by the marketer.

The fifth problem is that ROI has become a fashionable term without specific meaning. The 2005 American Marketing Association (AMA)/Aprimo, Inc. marketing accountability study identifies six ROI measures currently used:

Not one of these is actually ROI (although the second one is not far away). So it seems that marketers rarely mean "ROI" when they say "ROI."

The sixth, and possibly most serious, problem is ignoring the effect on the marketing asset (brand equity) that is a proxy for the longer term. Recall the AMA definition of marketing accountability quoted earlier. If marketing activities have generated $1 million in incremental profit after marketing costs of $500,000, ROI enthusiasts would applaud. If, however, those same marketing activities happened to simultaneously reduce the value of the marketing asset by $2 million, the story is reversed. This example underlines the importance, noted earlier, of using two types of measures to assess marketing performance: short-term profit or cash flow and the change in brand equity. In practice, ROI assesses only one.

Conclusions 
We need to separate performance evaluation from the comparison of future alternative marketing campaigns or planning. NPV techniques can be valuable for selecting strategy and planning. On the other hand, NPV techniques should not be used for performance evaluation. ROI promotes underperformance and short-termism, and should not be used in marketing measurement at all.

* Shareholder value can also be measured as the cash return to shareholders plus the increase in share value, but changes in share values are commonly held to anticipate future cash flows.

 

Adapted from London Business School working paper 05-709: "Choosing Marketing Dashboard Metrics" by Tim Ambler and John Roberts.

Tim Ambler is Senior Fellow at London Business School and author of Marketing and the Bottom Line (2nd edition – Financial Times Prentice Hall, 2003).

Click here to read Part 2 of this article.