As a consultant and researcher in personalization, I have a vested interest in asserting that personalization has immense value to an organization.
But flabby assertions don’t help you understand what to invest, or the return on existing in-vestments, or progress toward a goal, or setting the goal to begin with.
The leaders in personalization measure, track, and report results of personalization efforts. The challenges in doing so are both technical and mathematical.
Continuous multivariate testing efforts, which compare different approaches, discover and measure the optimum techniques deployed for personalization. The technical challenges are in deploying the various alternatives: identifying the best segment to experience each variant: cor-rectly assigning each visitor to a segment; and tracking, analyzing and reporting the results.
The mathematical challenge is evaluating the results of continuous optimization.
The so-called winner of each test is in turn nested into approaches that are tested and com-pared. After a year of continuous improvement efforts, how should the impact be calculated? Today’s revenue compared to one year ago? You’ll get a lot of argument on that one. Too many other factors: the sales team, brand marketing, the economy, competition, the weather, world events.
Successful personalization programs solve this puzzle by developing models that estimate impact and assign value. The models consider trends in the influencing factors, as well as actual visitor activity, extrapolating on where revenue (or other goals) would be with and without the program efforts.
Of course, models should always be viewed with a healthy dose of skepticism. While the algorithms may be sophisticated, the results can’t be terribly accurate. It’s a guess, not a measurement. But whatever the accuracy, having a consistent valuation method helps an organization proceed with its efforts and investments. Leaders in this arena have no difficulty justifying continued, and increased, investment in personalization.