Editor’s note: Michael Lieberman is founder and president of Multivariate Solutions, a statistical and market research consulting firm.

Over the past few years the awareness of totally unduplicated reach and frequency (TURF) analysis has grown. In fact, this useful, powerful technique has become more sophisticated, developing into more and more advanced applications than were originally envisioned when the technique first emerged in the 1970s.

TURF is often used to answer such questions as, “Of the 50 magazines on my list, which combination of five have the largest readership?” or “I have a budget of $10,000 - what are the optimal media outlets I should buy in?” or “What kind of market share will we gain if we add a new line to our current brand?” or “Which 10 of 20 flavors should my ice cream shop display?”

In the following piece we are going to present the most popular uses of TURF, such as maximizing reach and minimizing costs for media markets; calculating the incremental value to the full line of adding additional possible products; attracting the largest number of consumers with the fewest number of varieties; and projecting budgetary choices. An elegant upshot of TURF is that it not only allows the researcher to assess all possible combinations of brands or products on the list, but also points to the winners. In TURF literature, this process of determining the winners is referred to as optimization.

Finally, we will have a look at TURF deliverables, including ranked results (optimization), budgetary forecasts, and a TURF simulator that our clients find useful. It allows them to combine any what-if scenarios they like, or to pass it along to their client for the same purpose.

Components of TURF

TURF has two named built-in components, reach and frequency. In effective research, these two these pieces are generally separated, though they are often linked and both are universally presented with the final results. Understanding their simple meanings is a necessary prerequisite for understanding TURF.

Basically, frequency is the total number of people who will choose or purchase each product shown. For example, of 100 people 50 will purchase Sports Illustrated. The frequency for Sports Illustrated is 50. Sixty will purchase People magazine (they could also be the same people purchasing Sports Illustrated). People ’s frequency is 60.

Reach is the number of people who will purchase at least one of the products shown. Said another way, reach displays the unduplicated (no person is counted twice) percentage of people who will choose a combination. For example, Sports Illustrated has a frequency of 50, People has a frequency of 60, and Time has a frequency of 40. Out of the hundred people, 80 have read at least one of these magazines. That is the reach.

Figure 1 illustrates a simple example of an ice cream shop that only has room for three flavors but currently has six in stock. Respondents were asked, on a five-point scale, purchase intent for any of the six flavors shown. Top two box purchase intent is then calculated and shown as percentages. The frequencies presented illustrate the simple favorites of this store’s customers.

Which combinations of three, though, would tickle the tongues of most of the customers? Shown in Figure 2, we see that reach drops off by a few percentage points if you substitute caramel mocha fudge with French vanilla. Peanut butter fudge chunk, well, it just doesn’t add much to the mix, even if combined with French vanilla and caramel mocha fudge. The ice cream store owner has his answer.

Variation on theme

In a disguised example, Happy Detergent wished to accomplish two goals in one TURF analysis. Two measurements were involved to not only determine which combinations of products would sell better, but which of two package designs would be more appealing to consumers. The latter is derived by assessing reach percentages.

In addition, there was a caveat. Four of the brands were “fixed,” meaning that Happy Detergent already offered these products, and that fact wasn’t going to change despite the upcoming research results. The aim of the project was to measure the four fixed brands (perhaps for future phasing out), determine which new package works the best, and which additional brand would improve market share the most.

Which will be part of the line (fixed brands in bold)?

  • Happy Regular – Stain Dissolving Power
  • Happy Dazzling Whites and Colors
  • Happy Morning Breeze
  • Happy High Efficiency
  • Happy Sport
  • Happy Fragrance Free

A follow-up measure was asked, yielding a far richer lode of information. The two questions were:

  • Looking at all of these Happy brands, please tell me which statement on this card best describes how likely you would be to buy each of the specific types of Happy detergent you see here if they were available where you regularly shop? [for each package]
  • Assuming that (Happy brand) was available in the store where you shop, how many of your next 10 purchases of laundry detergent would be for this particular variety of Happy?

We see from Figure 3 that there are large differences (as shown by the frequencies) between perceptions of the Green Clover and Blue Diamond Packages. Which, though, is more effective? Figure 4 (reach) gives the answer.

So, we have a winner. Happy Detergent should go with the Blue Diamond. Looking at projected market share (Figure 5), Happy Sport would get around seven (6.9) of the next 10 purchases among this consumer sub-group.

If for some reason, however, Happy decided to go with the Green Clover package (perhaps it is less expensive to produce), our TURF analysis indicates that Happy would be better off with Happy Fragrance Free than Happy Sport.

Budget reach - a TURF pricing model

The next problem has to do with two questions. The first, “What is the best bang for the buck?” The second, “What is the best buck for the bang?” Rephrased, the first would be, “What is the best reach for a budget of, say, $5,000?” The second, “What is the minimum cost to get a reach of 80 percent?”

To illustrate, I will use a simple mix of women’s publications in which a client may potentially advertise. As part of the data-gathering process an online survey asks readers to which publication they currently subscribe. They are listed in Table 1 along with the costs to place an ad in each in measured units.

Let’s say the goal is for a certain reach. For example, if you had to reach 85 percent of the audience, what is the least expensive configuration of ads (Table 2)?

In our TURF software we would enter only one parameter, projected reach, then calculate the cost of scenarios that match. In our example above, where we are looking for 85 percent reach, two buy scenarios have been returned. One has a slightly higher reach, but the second buy, with four less-prestigious publications rather than three higher-priced magazines, is less expensive. Perhaps it is a better buy.

The second method of optimizing budget would be to enter in a cost. Using our example, say we would like to spend $5,000, and each $500 is worth one point. So, we are looking to spend 10 points (Table 3).

Here we again have two scenarios that emerge. However, the second, with two major magazines, has less of a reach than the first. So, we go with Vogue-Marie Claire-Essence. Again, the only parameter entered in the model is the amount of money the client would like to spend. TURF calculates the reach for all emerging choices.

Assessing the myriad

It is not hard to select three out of six ice cream flavors or five of six Happy brands, but what if the task was to select between the top 10 to 20 combinations of three or four options packages among 25. In our hypothetical example an emerging Internet broadband company (which we will call Netsurf) would like to test 35 options among broadband Internet users. Below is the survey question:

“Suppose (OFFER) was offered to new Netsurf for broadband.  Please tell me how likely you think you would be to subscribe to Netsurf for Broadband service in the next six months if this special offer were available?”

If we are looking for three out of 35, there would be 6,545 combinations to assess. If we are looking for four, there would be 52,360 possible combinations. If Netsurf wanted 10 out of 35, we would need to assess 183,579,396 different reach percentages. Listing out each and every one would drive the client crazy.

These days, though, it is simple (given large amounts of computing power) to run our proprietary program and calculate all 52,360 combinations of four (for example). Each is given a label, sorted high-to-low, then presented with the frequency and reach. We can test as many combinations as the client would like to assess.

Would a combination of the top frequencies receive the highest reach if three, or four, offers were publicized in one campaign? Figure 6 examines the top combinations and shows that TURF reveals that reach can often emerge differently than raw frequencies. That is, the top three and four combinations are not necessarily the top ranked individual offers.

Finally, another feature that clients enjoy is a TURF simulator, provided in Microsoft Excel, that allows the client to plug in any combination of offers to see what the reach might be. These do not necessarily need to be among the top scorers. It is possible that Netsurf would like to see the difference between certain middle-scoring offers, to test the drop-off if a higher-scoring combination makes a change, or to take the highest-scoring four-offer combination and see what adding other offers affects reach.

Figure 7 shows an abbreviated version of the Netsurf simulator (it would be difficult to show all 35 attributes within the confines of Figure 7). However, please note the “1” indicating the included offer and the “0” next to ones that are not shown in the current options. Also included in the simulator is each frequency for individual offer.

Value is added

These days brand managers are looking for the edge. They are turning more and more to the marriage of good research and advertising know-how to get there. The above mix of TURF methods - whether it includes brands, products, net profit, or total dollar sales - is the process by which value is added to the client/agency relationship.