ROI

Measuring return on investment (ROI) in marketing is an important and complicated issue. As with any other investment, businesses want to know they are getting value and increasing their bottom line. However, there are some fundamental differences between marketing and financial investments.

Marketing ROI isn’t straightforward. Different companies have different needs. Some sell expensive durable goods that are bought infrequently. Others sell packaged goods that can fly off the shelves. Still others sell business services with high acquisition costs and need to build long term relationships; the initial sale isn’t worth much. In fact new customers often have negative initial economic value.

So it is not surprising that evaluating ROI is confusing. Approaches can vary from an assistant with an excel sheet to sophisiticated software and high priced consultants.

The most important thing to remember about Marketing ROI is that there is no perfect formula. You have to choose the method that best fits your needs. A good place to start is to clarify what you want to achieve.

Company Strategy

How a business approaches ROI will depend a great deal on company strategy. Harvard professor Michael Porter lists three types of strategies:

Cost Leadership: Companies with a cost leadership strategy tend to be more focused on direct response advertising, so the ROI picture is somewhat clearer. Direct Response is fairly easy to measure.

Focus: Companies that focus on a particular market segment usually have limited marketing budgets. They are either regionally focused or cater to a particular type of customer. The marketing ROI equation is somewhat simplified for companies with a focused strategy. They don’t have that much going on.

Differentiation: Companies with differentiation strategies are the most marketing intensive. Their success depends on consumers believing that they offer something better than their competitors do.

Brands that seek to differentiate themselves need to build passion and desire for their brand. This strategy can be immensely profitable because consumers are willing to pay a premium for brands that they love and trust. Strong brands also tend to have lower acquisition costs.

For companies with a differentiation strategy, marketing ROI doesn’t just involve sales, but the perception of their brand that leads to sales. They track not only brand awareness, but also “brand attributes” – what people think about their brand. For instance, McDonald’s tracks attributes such as perception of value, quality, taste, “good for kids”, etc.

Marketing ROI for a differentiation strategy is complicated further if it involves a durable good with a long product cycle (i.e. cars, home appliances, etc.). Consumers who desire their product might be years away from a purchase decision.

Four Basic Approaches to Marketing ROI

Timelines: Making a timeline of marketing actions can be very useful. You can see what you did, when you did it and track response, such as sales, inquiries to the call center, etc. By comparing marketing inputs and outputs you can get a sense of what drives sales. If you can identify a correlation, the math is pretty straightforward and you can get a good idea of what drives your ROI.

While the approach has the virtue of simplicity, it also has its problems. For instance, a timeline only shows when the action happened, not how intense or how long it was. Furthermore, if there is a lot of activity, the timeline becomes unreadable and loses the virtue of clarity.

Grids: Because of the limitations of timelines, many marketers use grids. They are very easy to make in excel and you can overlay activities of different durations.

Moreover, additional information can be inserted into the grid. For instance, you can show that you bought 200 GRP’s of TV per week for 5 weeks and also ten monthly magazine insertions and 1 million banner impressions per week for six weeks. Grids retain much of the simplicity of timelines and contain much more information and flexibility.

However, grids too have their drawbacks. How should intensity be measured? For TV, is GRP the right metric to use, or should it be coverage at a discreet frequency or share of spend in the category? For internet, should banner shows be in the grid? Clicks? Registrations?

Grids also can also get cluttered and confusing. For some very active brands, they become enormous and unwieldy.

Tracking and Two-Variable Modeling: Another approach is to track all activity using a variety of metrics. With tracking, you lose the visual simplicity of grids, but you can include as much data as you want.

You don’t have to guess beforehand what you think will be most relevant. After tracking for six months or so, you should have enough data to build two-variable econometric models. (For more on models, see Less Numbers – More Math).

Two variable models require some mathematical sophistication, but can be done in excel. No expensive software is required. There are some subjective decisions to be made about which model to fit, but common sense goes a long way. Usually the simplest model is best.

Once you find a good model, you can derive the equation and significantly increase efficiency through better planning. However, two variable modeling can be very labor intensive. You need to try a number of variants before you find one that fits well enough to be useful.

Unfortunately, if there is too much market activity, it will be hard to find a two-variable model that explains more than 50% of the variation in the data. A model that offers more doubt than certainty is of questionable value.

Multivariate Modeling: Some of the world’s premier marketers use multivariate modeling. While prohibitively expensive for most companies, you can include a variety of marketing and economic factors and get much better models than you can using only two-variables.

However, multivariate models are extremely complex. PhD level mathematicians need to be contracted. Moreover, designing a good model is as much art as it is science. Those that build the model have to understand the business and the business people who use the model need some understanding of how it works.

For a successful multivariate effort, the rare combination of high level quantitative skills and street level business acumen is essential. That’s difficult to achieve.

Furthermore, even the best models will fail at some point. As markets evolve, success factors change. The world is a messy place.

Beware of Easy Answers

Often, the story is more complicated than it seems at first. Years ago, I had a magazine consulting client who was experiencing a fall in copy sales. Because there were no new launches in their category and they had no change in their marketing strategy they assumed that their product needed to be reworked.

It turned out that, although they hadn’t altered their marketing intensity, competitive activity had doubled. Comparatively, my client’s marketing activity had dropped in half. By raising their marketing budget to meet the competition, they were able to reverse the slide and maintain the integrity of their product.

It’s because of the complexities described above that ROI is probably the most difficult and contentious issue in the marketing arena today. There are many ways to get it wrong and very few ways to get it right.

Some Common Sense ROI Rules to Follow

Keep it as simple as you can: One core principle of econometrics is to always use the simplest model that explains the data adequately. If you’re getting reasonably good insight with a simple process, there is no point in trying to be more sophisticated.

Define your goals: No ROI process can explain everything. While a perfume brand might want to focus on consumer perception a retail business will focus more on direct sales. You can’t optimize your marketing strategy for everything. Managing a business is about making choices.

Check your work: Once you believe that you have identified a factor that drives your marketing performance, try to get the same result using a different method. When evaluating marketing ROI, it’s easy to get false positives.

Avoid bias: If at all possible, people who are implementing campaigns shouldn’t be responsible for evaluating them. Suppliers, for the most part, shouldn’t even bother with ROI. Unless there is a very integrated relationship with the client the analysis will be ignored at best. At worst, it can be selectively used against you.

Eventually your model will fail: Even the best models are based on past experience and any bearing on future performance is tentative. A failed model shouldn’t necessarily be seen as a bad thing. It is a signal that the basis of competition has changed, and that is very important for a marketer to know.

I hope this has been helpful.