The Folk Apostle Model
Posted by Jeffrey Henning on Sun, Aug 17, 2008
The classic Apostle Model was not actually graphed but was simply described in a table. Later users of the model did graph it, but they often oversimplified it as a quadrant analysis. The following illustration is typical of the type of analysis I've seen from many different companies:
This analysis is wrong, as Jones and Sasser group "low" and "medium" together, so that "high" for satisfaction and loyalty should represent only one third of the grid (see yesterday's post). Why did the authors specify this? Because ratings of a 3 to 5 on a five-point scale do not indicate high satisfaction or loyalty; only a rating of 5 itself does. For instance, Jones and Sasser report that for Xerox completely satisfied customers (rating of 5) were six times more likely to repurchase over the next 18 months than somewhat satisfied customers (ratings of 3-4). This distinction between high scores and medium scores has also been demonstrated by Reichheld with the Net Promoter Score, where only ratings of 9-10 on a 0-10 scale indicate a high result (roughly comparable to the 5 on a 5-point scale).
Another common mistake is to use four-point scales. One of our competitors recommends that loyalty be measured with a repurchase-likelihood scale of "Definitely Will Not, Probably Will Not, Probably Will, Definitely Will". As Brad Bortner argues in his recent Forrester white paper, a scale with too few points "lacks enough granularity". I'll recommend some best practices for scales for the Apostle model later in the week.
I would imagine that this common mistaken use of the Apostle Model started for three reasons:
- The original authors didn't include their own illustration, which would have served as a template for later users of the model.
- Many consultants are familiar with different types of quadrant analysis and simply adapted that common practice to the Apostle Model.
- Many firms found that they had more Apostles when they inadvertently "cheated" at the customer satisfaction graph than when they did it correctly.
Tomorrow, I'll turn to the Prisoner/Apostle Model, an adaption of the Apostle Model for service businesses.
Update: This post is part of the series The Apostle Model and Related Loyalty Segmentations.