Data from a scientific test are like pieces of a jigsaw puzzle. As the test goes along, you get hints of which elements may be helping or hurting. Then when your final data are analyzed, all the pieces come together and you see the real-world impact of every test element on its own and in combination with others.

The sign of a well-run test is simple and powerful conclusions

A skilled test expert can summarize the complexity of analyses into simple, clear graphs and charts. As data are analyzed, the depth of marketing insights grows.

Skilled analyses give deeper insights into the true impact of the test elements: 1. Test recipes (like split-run test cells) show which combinations are doing well, but not why. 2. Main effects quantify the real-world impact of each test element, showing which help, hurt, and make no difference (as the case studies show, all effects should be summarized in a bar chart). 3. Interactions give deeper, more accurate insights into how the main effects vary depending on how other elements are set (see interaction plots in the case studies & articles) 4. Test the optimal combination versus the original control and calculate your ROI The optimal combination is created by implementing all new ideas that increase sales, avoiding changes that hurt, and keeping the control (or the lower-cost) level for all non-significant element. The optimal combination often performs much better than any test recipe, since the best combination is seldom one of the test recipes. Based on these insights, plan further testing. This one test may not answer all of your questions, or results may lead you in a new direction. This is why testing should be an on-going process. You learn a great deal in one test and then leverage those insights to decide what to implement and what to test further. A well-planned and well-executed test will provide clear answers that your marketing team can understand. However, if your sample size was too small, test elements weak, recipes executed inconsistently, or data quality poor, then you may find confusing data, unclear effects, or few significant effects.

Like a virtuoso violinist or skilled copywriter, testing looks so easy when done well

All your effort defining test elements, creating recipes, and launching the test will be rewarded with clear, actionable results. Surprisingly, the complex techniques behind scientific testing are what give you deeper, more accurate—yet straightforward—insights into the true impact and relationships among your test elements.

In some ways, in-market testing is the mirror image of standard analytical techniques. Most statistics—like data-mining, modeling, neural networks, and discriminant analysis—take readily-available data and look for important correlations among the variables. However, the lack of structured data requires very complex analyses to uncover logical relationships among a mess of random data points.

Testing, on the other hand, requires careful up-front planning and design. Proving what works in a complex and dynamic marketplace is never easy, but proactive scientific testing gives you the freedom to design your marketing programs to give you the information you want, quickly and clearly. The better the test, the easier it is to understand and implement the results. Numerous examples are shown in the case studies & articles.

A well-run scientific test can give you an impressive ROI with amazingly clear results. The final step of the LucidView Strategy is profit.

Back to top