I made it sound like some scientific procedure, obviously it’s not but the benefits are multifold :
- Is my planning and strategy in the right direction ?
- Which campaign format will get me maximum visibility and conversion ?
- Which channel should I be focusing on more ?
- What is the expected ROI that I can forecast ?
- Can the ROI expectation be distributed between different channels ?
Your team might have multiple opinions on how the campaign should be designed, where the banners should be placed, the flash animation, left navigation, colors, messaging, and many more. All of them might be good and rightly so, creativity has no rights or wrongs. Testing campaigns takes away the ambiguity of eliminating and retaining the opinions/advices based on maximum benefits.
We might call it A/B testing, MVT, Test and Target, etc. In simple terms, create multiple versions of your campaign with all the creativity into consideration, create customer segments you would like to target and launch all the different versions. You might do it in different ways, there’s no protocol here. I can tell you what really worked well for me.
We created 5 versions of our campaign with different designs. Creativity all right !! but there has to be a limit, right ?. Test plan was created 4 weeks prior to the launch; the timelines also depends on how extensive the campaign is, if it’s small scale, then may be a week of testing should be enough. Divide your target region into 5 sections and launch your versions. Measure real time, you might get quick and clear results. Initially we planned that we’ll run and measure all the 5 versions for 2 weeks but the results were so “on our faces” that a week was enough. We Eliminated the bottom 3 campaigns and divided the target region again into two this time. Same procedure again and we were left with our best campaign.
Remember the best campaign can be a mix of your 2-3 versions as well. If you measure the results by sections : banner, masterhead, footer, leftnav etc, then pick up and Retain what’s best for all these sections from all the versions, you get best of all. No, I will not ask you to test it again, let’s not stretch it.
Testing campaigns on different channels is also an option and it can help you determine your ROI expected from different channels. But in today’s time and age customer doesn’t use a single channel to purchase. For them a channel is an act of convenience and experience. Jack will not buy a sofa just by looking at the website. He’ll check the pictures, like it on Facebook, tweet his friends and ask for their opinion, check from influencers through blogs and forums, visit the store to touch and feel and then finally visit the website again to buy it. So long story short, I am always in favor of testing the campaigns on the website and external linkages like Amazon which land on your website. If it doesn’t work here, it’s bound to bust everywhere.
Finally I hope by now you have all your metrics in place, measure them like you would measure your real campaigns and formulate your revenue based metrics into a ROI. It will be interesting to know how well you estimated when you run your real campaign. Mine was pretty close.
Next Up : Measuring the metrics in the most important and result oriented aspect of a campaign. There are several metrics but do you want a longlist of all these metrics on your table every morning ? We’ll talk about how you can determine if the campaign is on track by just reviewing 5 metrics. These are the ones which can raise alarms or open champagne bottles.