Marketing is often said to be all about getting clients. While many newer marketers tweak this a little to remind everyone that it is also about getting the most from the business to client relationship, still the number one concern is in finding business to service. Split testing in online internet marketing or offline direct response marketing is key in doing two things to help this:
In short, split testing is a technique for testing the copy, measuring the response rate, and feeding that information back into the loop to be re-tested. In this way, we can narrow the focus of our marketing effort to achieve stunning results.
We need to be able to measure our efforts. Typically, for an overall campaign, we would like to know:
There are more variables, but these are among the most important. Online, it is easy to track the conversion rate, as we can measure the web statistics. Offline, we need to remember to ask the customer to quote a specific offer number. Once we can track an offer, the rest of the values can be measured. We can not test anything that we do not measure.
An A/B split test is very simple. We create two versions of our copy, and change one part of the piece:
We then funnel half the leads to each, and measure the conversion rate based on the offer. We now know which is best. If both conversion rates are the same then either the two pieces were still too similar, or both choices were of equal merit.
So, we take one, split and test again with a third value. This is best illustrated with price : if we start off with a low price point and a medium price point, and find that the conversion rate is higher for the mid price, we take that and a high price point and test again.
Taking the price split example in internet marketing, we can generate two pages. In fact, the flexibility of the internet allows us to generate the price and offer identification dynamically, based on either a 50/50 split, or dynamically depending on ongoing results.
Offline it is a little more complex, but we can still perform the test on, say, 10% of a given list of leads. The goal is the same : find the point at which one offer significantly outperforms the other, and use that as the control piece for a test mailing or online campaign.
We can split by offer type also; using a hard offer (send us money and we'll refund you) versus a soft offer (send no money now, we'll bill you later). Some products attract customers that respond better to a soft offer, and others a hard offer. The only way to know for sure is to use testing, and the easiest to pull off is split testing.
So, the next time a marketing effort is being put together, it might be worthwhile to ask the copywriter to produce two versions of the key points of the piece (headline, closing, lead paragraph, etc.) and work on finding the killer combination.
The technique can also be applied to display adverts and other forms of marketing and advertising : just remember to have a way of tracking the conversion rate per split, otherwise the effort is lost.