Squeezing More Revenue From Your Marketing Dollars

This post from Paul Gorman originally appeared in March 2015 on Renaissance Executive Forums.

Marketing pioneer, John Wanamaker, once said “Half the marketing money I spend on advertising is wasted; the trouble is I just don’t know which half.”

I’m not going to tell you how to completely and totally fix that problem, but there is a way to significantly increase your revenue and, at the same time, reduce how much you spend on marketing and direct mail. We can do that through the power of the computer, big data and predictive analysis.

What Is Predictive Analysis?

We all know what big data is and you probably have an enormous amount of it stored on your computers or in an archive. Predictive analysis mines that data, turning it into information that can be acted on. By mining the historical data that you keep collecting, we can pull out information from the data and find statistical relationships and meaning that best describe the past behavior of your customers. We then apply that knowledge to predict future customer behavior and improve the bottom line of your business.

Probably the best known application of predictive analysis is your own FICO score, a measure of consumer credit risk. Your FICO score rates you as a consumer. Almost everybody has a FICO score even if they are not aware that they do. It gives your lender a way to predict the probability that you will default on your loan and the likelihood that he will get his money back. The higher your score, the less likely you are to default on your debt, and, therefore, you represent less of a risk to the lender. As a result, with a high FICO score you can typically qualify for better terms when you buy a house or car, or apply for any other kind of loan. With a low FICO score you may be able to obtain a loan, but it may be financially painful for you because you’ll have to pay a premium to make up for the lender’s risk.

Where Is Predictive Analysis Applied?

Predictive analysis can be applied to a myriad of fields besides credit, marketing and direct mail. One example of where predictive analysis is paying big dividends is law enforcement. For instance, the San Diego County Sheriff has a small staff of analysts applying predictive analysis in order to help reduce jail overcrowding as a result of recidivism. This not only reduces jail overcrowding, but reduces taxpayer expenses as well.

A few other examples include: stock market trading; safety and failure analysis; employee retention; fraud detection; network intrusion detection; and political campaigning with voter persuasion modeling.

How Can Predictive Analysis Help Your Business?

Direct mail marketing is one important area where predictive analysis excels and can pay big dividends to your business. In this case, after mining the data and applying our predictive analysis techniques, we can assign probabilities or scores to each individual in your database, and can rank them according to the likelihood of their behavior or actions. Now we can easily tell who your best customers are, ranked from top to bottom. This is an objective measure, not subjective, and allows us to separate the “wheat from the chaff.” Subjectively though, you still have the option of including Uncle Bill in the list if you want to, even though he is way down in the bottom half, and just loves the free buffet.

If we already have a direct mail list, we can then match that list against what we now know. This produces a more optimal mailing list, one typically far superior to the one created in almost any other way. For instance, if we have a million people in our database, but are mailing to only half of them, we can now be very sure that we are mailing to only the top half. The cream has risen to the top. (I love mixing metaphors.)

Once we have that information, we can then even further refine our superior mailing list and begin to predict what each customer is most likely to respond to. That way, instead of just shot-gunning one or two offers to all half million of our best customers, we can actually segment them into categories, and tailor our marketing much more precisely to appeal to each segment. The best number of segments and types of offers can also be found through predictive analysis.

Altogether, the results of this approach are to (a.) market to those customers whom we know are our best customers; (b.) present tailored offers to those customers in a way that they are most likely to respond; and (c.) actually take control of our direct marketing campaigns and dollars by choosing not only who to market to, but also who not to market to. As hard as not mailing to someone may be – such as Uncle Bill – we can now objectively know that the customers way down our mailing list are probably just not worth the stamp. This, of course, reduces marketing costs.

Example 1 – Mailing Better Customers

Here is an example using our database of 1 million people. Let’s assume that that you mail to half of them, and it costs $2 to mail to each one. You have observed that your response rate is 1%, and the average revenue produced by each customer who responds is $200.

Profit = Revenue – Cost

Profit = $1,000,000 – $1,000,000 = $0

So, why are we doing direct mail again?

Predictive analysis allows us to improve on these results. Now suppose that your predictive analyst crunches some numbers, and is able to rank your customer database from top to bottom. He finds that he can replace half of your original mailing list with better customers. As a result, your response rate increases to 1.5%. Also, because you are mailing to an optimal set of customers, you can expect that the average revenue produced jumps to $225. Now let’s see what happens.

Profit = Revenue – Cost

Profit = $1,687,500 – $1,000,000 = $687,500

Example 2 – Mail To Only The Best

The second way to improve bottom line results is to reduce the size of our mailing list, and save the money. This, of course, cuts costs. In this example, if we only mail to the top 250,000 customers, then we will have saved $500,000 in mailing costs. But that comes at a price. Since the customers we are no longer mailing to are at the bottom of the list, their response rate and average revenue is less than for the entire group of 500,000. Therefore, let’s assume that the response rate for those included in the mailing increases a little, to say 1.6% and our average revenue increases a little to $230. Let’s see what happens.

Profit = Revenue – Cost

Profit = $920,000 – $500,000 = $420,000

In this example, we probably cut our mailing list size too much, but you get the point. We saved $500,000 in costs, but only reduced our resulting revenue by $267,500. Nevertheless, we may have gone too far in our cost-cutting.

But what if you only have a $300,000 direct mail budget? Then that additional $267,500 starts to look pretty good. How can you start to make up the difference on the revenue side?

Example 3 – Tailor Offers to Value

One way to raise revenue and still keep our direct mail size and costs down is to be smarter about what we offer our customers. As mentioned above, we can better tailor our offers to better fit specific segments of our mailing list. This will improve our response rate and the average revenue per customer will increase as well. In our simple example above, improving the response rate by only 0.25% increases our revenue by about $300,000 while our mailing costs remain the same.

Here is how this could work. As a strategy, let’s still use the $500,000 direct mail budget, but segment our 250,000 mail list in to 2 groups of 125,000 each. In the top group we can expect a response rate of 1.75% and in the lower half a response rate of 1.5%. In the top group we can expect an average revenue of $250, and in the lower group average revenue of $230. Now what happens?

Profit = Revenue – Cost

Profit = ($546,875 + 431,250) – $500,000 = $478,125

So compared to our results in Example 2, we have increased our expected profit by $58,125 by minimally segmenting our database and sending two different offers, each more aligned with the needs and interests of our customers. We can probably improve even more on our results by further segmentation and tailoring.

The Bottom Line

Now direct mail starts to make sense again, and we have a way to squeeze more revenue from our marketing dollars.

Let’s recap and look at where we’ve actually come from when we started this discussion. We started with a huge mailing list, an OK response rate, an OK average revenue per customer, but zero profit from our efforts. By applying Predictive Analysis step by step through a rigorous process, we’ve actually improved both our response rate and our average per customer revenue while we reduced the size of our mailing list and cut costs.

Increased revenue. Lower costs. What’s not to like?