Predictive Marketing Blog

How to Find Your Best Prospects Using Analytics – Part 2

It is imperative for marketers to efficiently use their advertising dollars in order to effectively reach the most likely prospects to buy their products or services in order to grow their businesses profitably.

In part 1 of this series we examined a direct marketing campaign generating 17,000 sales and identified the demographic target of these new customers based solely on identifying the highest percentage demographic factors from a market research study of these new customers; we found that this targeting approach significantly underperformed a random set of prospects who received the same offer. Part 2 will explain a basic approach of how to find your right prospect target.

What is the right process for marketers to use to identify their true likely new prospect characteristics?

In part 1, we found that employment, 3-5 years at their residence, and age under 65, identified 50% or more of the new customers from the market research study. Using the campaign database of 17,000 new customers, that also contained a 200,000 record random sample of the advertised market, we analyzed all 75 combinations of the 3 major market research profile variables (occupation, age, and length at current residence) to see where new customers differed from the advertised market. By computing the percentages of buyers to the percentages of the random market for all the 75 combinations of occupation, age, and length of residence, we can identify segments that have high percentage rates for customers and lower percentage rates for the random market.

We looked within the ‘employed’ segment and examined the percentages to see where new customers were strongly different from the market; where these characteristics can identify better prospect names for marketing.

We found those who are employed and are new residents (1 year) at their address and are older (ages 65+) had greater concentrations in new customers than the market percentages. This finding is opposite the initial market research approach; in fact, we found in this segment (see figure #3), that they are 68% better than random and 118% better than the erroneously constructed market research percentage.

Figure #3: We examined the random database to determine the sale rates of those who were employed to see what sub-characteristics identify buyers that are better than random buyers.

Fig3

Marketing’s objective is to identify and then advertise to prospects who are most likely to become new customers. For advertising to be cost effective the marketer must create the maximum number of sales for a minimum cost in advertising. This means identifying demographic characteristics that differentiate new buyers and where they differ most from the market. Collecting market research demographics on new customers in a vacuum is not guaranteed to identify characteristics of the marketing universe that isolate new customers from the general market. Marketers need to look for demographic characteristics and their values where their new customers have higher concentration or penetration rates than found the market in general.

Using this penetration rate approach (perhaps called a matrix or interaction approach), a new buyer profile is revealed. In our example database, the best matrix target is now made up of an entirely different set of values:

1. Retired (vs. employed)
2. Length of residence 1 year (vs. 3-5 years)
3. Senior, 65+ aged household (vs. under 65)

This set of values is quite different from the original approach and it a basic target market.

Figure #4: Using the same 3 demographics from the market research method and employing the penetration* index analysis to identify a profile of best buyer prospects for marketing, we find:

Fig4

When the sales rate of the penetration index approach is calculated (fig. #5) we see a stronger prospect target emerging.

Figure #5: Using the random advertising database to determine the sale rates of those who were identified by penetration matrix to be best prospects, we see they are better than random.

Fig5

 

While the penetration index approach works well and clearly is better than the market research approach, it can be fairly unwieldy and not the optimal approach either. This new segment only represents 2% of the random market to target.

 

Conclusion

This penetration index approach is accurate and usable when only examining a few factors but becomes unmanageable when there are more than a few main demographics. Imagine manually manipulating hundreds of variables and thousands of values in a matrix; it would be nearly impossible. And given the big data available for analyses today you would be shortchanging your targeting effectiveness and efficiency by limiting your matrix to only a few characteristics.

Big data analysis is needed to clearly identify:
• the most prospects
• with the best penetration rates
This big data approach to targeting uses statistical analysis to define the best target market.

In part 3 of this series, we’ll discuss the statistical analysis approach to identify who marketing should be targeting and how to easily identify your best new customer prospects.