With TESTING, the single-word mantra for any one-to-one marketer, mistakes can be corrected and bugs can be ironed out before thousands of dollars (replace with’ lakhs of rupees?’ ) are spent. Testing is a way of minimizing risk as well as maximizing results.
One-to-one marketing is all about tailor-made offers and relevant communication to one target individual or a focused group of target individuals. What differentiates it from mass communication is that all one-to-one communication is response oriented. Apart from the relatively broader objectives of mass communication in terms of awareness/perception measures, all direct communication is oriented towards a specific action or response – be it a write-back, call-back, log-on or even ‘purchase a trial pack.’
As agencies and as program managers, we often have the unenviable task of predicting ‘response rates’. The entire economics of a program or campaign often hinges on these predicted rates that are, more often than not, little more than thumb rules and guess work. This is where we come to the single word mantra for any one-to-one marketer – TESTING. When in doubt – test.
WHAT TO TEST?
Typically response rates in any campaign are affected by 5 key factors : (in order of priority)
| Factor Name |
Description |
Variation in response rate |
| LIST |
who we are talking to |
6: 1 |
| OFFER |
What is in it for the consumer? |
3:1 |
| TIMING |
When is the communication being sent out? |
2:1 |
| CREATIVE |
What is the content and form of communication? |
1.35 : 1 |
| RESPONSE DEVICE |
What is the instrument that invites the response? |
1.2 : 1 |
Therefore the difference between the best and worst response rates in any campaign can be as high as 58 times to 1!
HOW TO TEST?
A test plan needs to keep in mind the following:
-
Test different variables that may impact response: Traditionally, only one variable was tested keeping other factors constant. However, now we can test multiple variables at the same time to determine the greatest impact on response.
- Test samples need to be statistically significant. This means that we should be able to extrapolate the results we get to the entire database. To that extent e-mailing to 10 people and getting 6 responses does not mean a 60% response rate can be expected for the roll-out! There are specific formulae to arrive at statistically significant sample sizes.
- Test group results need to be compared to a matching control group. If the test group does better, then we say we have beaten the control and the new combination of list/offer etc is used for the rollout and becomes the new control.
- Any test plan must be given all chances of success. In this manner, we can be reasonably sure of whether it will work in a roll-out situation or not. Running a test plan in a historically weak market and failing does not mean that it would fail in a strong market too.
Testing has the following advantages:
- It helps the marketer ‘look’ before he ‘leaps’. Therefore it helps him gauge results on a smaller budget before committing all the monies upfront.
- It is an ongoing process of continuously fine-tuning and improving results reflecting a discontent with status quo. More benefit for the customer and better economics for the marketer; a win-win situation all the way!
Due to operational deadlines and constraints however, testing is often sidelined to make way for blind repetition of previous campaigns or short-cut, off-the-cuff campaigns to the entire base. While marketers and agencies have to live with the inevitable reality of deadlines and time constraints, it is important that every annual plan sets aside a ‘testing’ budget so that test modules are compulsorily conducted every year on one or more variables with a view to improving response rates.
|
