Thursday, August 28, 2008

Measuring the Value of a Marketing Measurement Project - Part 3 of 3

So far this series of posts has described generic inputs to calculate the value of a marketing measurement project. Ordinarily an internal expert or a consultant would help define the values for those inputs. But let’s assume we’re in the self-service world where non-experts provide those values for themselves. What advice can we give them?

The first is to define the project scope. That is, you want to limit the analysis to the customers, revenues and costs that will actually be affected by the project. For example, inputs related to a Web analytics project should only include the number of Web customers, Web revenues and Web costs. But even this case isn’t so simple: if the project is measuring something specific, such as paid search results, then it may affect just a subset of all Web customers. It’s easy to overestimate the scope of a project’s impact, which in turn overstates the expected benefit. The more precise you can be in your understanding of the actual customers affected by a project, the more you’ll be able to build a realistic estimate of its impact.

A second useful consideration is the actual mechanism that will provide the expected change. Because we are talking about marketing measurement projects, all the project provides is better information about business results. Unless this information can be used to change something, it won’t have any impact. So, for example, a system to uncover immediate opportunities is worthless if you can’t react in less than a week. Similarly, a model that shows the value of changes in the marketing mix has no value if the mix won’t be changed for political reasons. A close look at how the new information will be used will allow a more precise estimate of the expected changes in value.

A third consideration is your ability to project the impact of your changes. Most marketing measurement projects can show the relative performance of current marketing programs, but only some can estimate the impact of incremental investments in those programs. This can be a particular issue where there are constraints on program expansion, such as limited advertising inventory. Where projections of incremental returns are not immediately available, you may be able to conduct additional research or use standards rules of thumb such as the square root rule to make reasonable estimates.

Finally, you need to consider the time horizon of your analysis. Most MPM projects will require a significant initial investment followed by lower, recurring operating costs. The benefits will probably follow a very different pattern, starting slowly and then growing over time. The formulas I’ve presented can be extended to create this sort of projection by doing separate calculations for a sequence of time periods. But in many cases a much simpler approach is acceptable, basically of estimating the benefit for a “mature” program for a fixed time period such as one year. This can be compared with program investment to calculate annual ROI or pay-back period.

Note that even a “simple” calculation may need to consider long-term impacts. For example, a program that acquires new customers more effectively will benefit from the full lifetime value of those customers, not simply their first-year revenue. This is where the idea of “mature” value comes in. It implies looking at a slice of results across all customer groups after the program has been in place for some time. Doing this math correctly can be fairly complex. But, again, you may be able to save effort by making some simplifying assumptions. In the case of a program that let’s you spend more efficiently on acquisitions, this might mean assuming the number of new customers will remain the same because you reduce your acquisition budget. Since the number of new customers doesn’t change, the later values also remain the same. The value of the program thus equals the savings in acquisition cost—a much simpler item to estimate. (Of course, this will underestimate the actual program value if you will in fact be maintaining your acquisition budget and acquiring more customers. Whether this matters depends on the situation.)

There are certainly other considerations that could apply in a given situation. Identifying them all may not be possible. But even if you could list them, the resulting document would be too long to be practical in a self-service situation. So we have to hope that our self-service consumers are able to complete the process on their own, or at least recognize when it’s time to call for expert help.

No comments: