- 1 - Decide1 - Decide
- 2 - Run the models
- 3 - Launch the campaign
- 4 - Read and find out
1 - Decide
Time has come to unleash your creativity.
In this first step, you will design your test campaign by answering a number of questions around:
- Your strategic objective - why do you want to test (increase redemption, maximize profitability, ...)?
- Your test content - which product / service and which dimensions do you want to test (pricing, features, communication, promo, ...)?
- Your test structure - which size and complexity for your test (number of cells, cross-effects integration, ...)?
- Your target - on which customers do you want to test (representative sample, specific segments, ...)?
- Your channel - which contact mode do you want to use for your test and the possible follow-up (outbound call, mailing, ...)?
Example of test content: pricing and communication test on revolving cards.In this example, a leading credit card company wanted to optimize the profitability of its direct sales by identifying the best value proposition for different customer segments (mailing campaign). Five components of the value proposition had been considered as potentially impactful by the Marketing team, which wanted to "challenge" each of them during this Test & Learn campaign: maximum length for credit line reimbursement, APR value (interest rate), mailing "theme" (message), limited offer or not (and if yes, what kind of limit) and gift offer for subscribers (and if yes, what kind of gift). For each of this 5 dimensions, the Marketing team decided to test between 3 and 5 different values, for a total of 18 values... corresponding to 540 theoretical combinations! Needless to say that testing all of these combinations in order to find the best one would have been clearly impossible without a Test & Learn approach.
2 - Run the models
Science is an art. Use the right tools to get the best of it.
In this second step, you will design your final test structure, by focusing on 2 key elements:
- The Experimental Design, which will identify the actual test cells.
- The Sample Homogenization, which will distribute homogeneously your customers sample along the test cells.
Example of Experimental Design and Homogenization: pricing and communication test on revolving cards.In this example, the credit card company seen in the previous step ("Decide") managed to reduce considerably the number of different offers to be actually tested on the market: from 540 theoretical combinations to only 14 test cells. This has been done with Experimental Design models. You'll notice that the cell #6 has been highlighted within the Design ("forced inclusion"), since it is considered as the "standard" offer and will act as a benchmark to measure the potential upside of the best offers. For each single cell, a group of 20.000 customers has been idenditified (except for the "standard" cell which is comprised of 40.000 customers). These customers have been selected in order to ensure an almost perfect homogeneity between all test cells.
3 - Launch the campaign
Now launch your campaign. And follow it closely.
This step is mostly yours. No models, no surprise. You concretely build your Direct Marketing campaign as you would do for any other initiative in your marketing plan. The only difference is in the number of offers that you will send simultaneously: sometimes it may be only 4 to 5, sometimes you may reach 15 to 20 different communications / value propositions. It all depends on the complexity of your test. But, in any case, there is no big hurdle here - it's only a question of organization. Once your campaign is launched, your customers become the main protagonists. As in any Direct Marketing campaign, some will answer quickly and some will never do. But there's a big difference here as well, compared to a standard marketing campaign: their answers are directly linked to a specific offer and will tell you, for the first time, the truth about their actual preference among the infinite number of small variations of your value proposition. As usual, you will have to wait a bit before drawing any kind of conclusion regarding the success of your campaign: generally 2.5 to 3.5 months when talking about a mailing campaign for instance. But in the meantime, you have to keep a clear vision of your campaign's status. You may already have some kind of reporting tools to follow your Direct Marketing campaigns over time. In the case of Test & Learn, some additional elements might be useful if you want to anticipate some findings or, in some cases, adjust the campaigns characteristics (add one or more cells in the case of outbound calls for instance, ...).
Example of reporting tool.
4 - Read and find out
That's it. You know where to go.
In this fourth and last step, you will identify the best possible combination and assess its performance increase compared to the standard offer. 3 key actions have to be performed:
- 1. Analyze the final results
- 2. Extrapolate the results to all theoretical combinations
- 3. Identify your preferred combination, according to your strategic objectives
For the first time, you will be able to assess accurately the expected revenues and profitability of a value proposition that you never launched on the market.Once again, you can decide to do that for the whole population, or to differentiate your analysis per customer segment. Based on these findings, you will rank all your theoretical offers according to their performance around one or more of the 3 main indicators (redemption, revenues, profitability), and you will have all the elements to decide which offer(s) should be leveraged during the next marketing campaigns. Most of the time, there is more than one "best" offer: the one which maximizes your redemption is not necessarily the same that maximizes your profitability. And, obviously, you will end up with more than one best offer if you decided to go for a cluster-oriented approach. In any case, you will find the truth.