A / B Testing: Experimenting with one or more items at a time?
The month of August is a bit slow in terms of written articles, so I take advantage of the relative calm of this Saturday to write about a very interesting topic.
Carry out an agenda of experiments with different landing pages, creatives, images, calls to action, etc. It is an essential part of the optimization process, not only for Google Adwords campaigns, but for online marketing efforts in general.
There are also tools for all tastes, free such as Google Website Optimizer or you pay as Omniture Test & Target, so there are no excuses.
One of the most frequent questions is whether you should experiment with one or more elements of a landing page at the same time, or even if you should experiment with a totally different landing page.
The answer is almost the same as for all other questions ... it depends on your current goals.
When to experiment with a single element
Experimenting with a single element of a landing page (be it an image, call to action, form submit button, etc.) is great for having very precise control over how the change made had an impact on the performance of the campaign. , this is particularly VERY difficult when experiencing multiple modifications at the same time.
In the same way, if our campaign has a Return on Investment that according to our objectives can be considered very good, perhaps there is a small "fear" of totally ruining the performance of a campaign with a landing that may not work so well. It is popularly said "Do not fix what is not broken."
Experimenting with a single element reduces the chances that the performance of the campaign will deteriorate catastrophically, although it is also very likely that the performance will not improve very abruptly either, of course there are exceptions and very curious cases where changing only one element turns around the whole board.
In short, if things are going relatively well, I would only advise experimenting with very specific elements of the landing to gradually refine the Return on Investment obtained.
When to experiment with multiple elements or a totally different landing
Suppose we are getting a Conversion Rate of 1.5% when we really need at least 5% to make the campaign profitable, we are in trouble and the customer is not very happy.
In these cases, common sense dictates that changing a single element of the landing may not mean a significant change in performance, since it is most likely that SEVERAL elements of the landing are causing the performance problem.
All this assuming that the quality of the traffic is good, if the campaigns of Adwords, Email Marketing, Affiliates, etc. they are providing low quality traffic there is no landing or A / B Testing that will save the day… but that is the topic for another article.
When things are clearly not going well and drastic changes in campaign performance are needed to achieve our objectives, the ideal is to identify what are the possible elements that are causing the bad performance to create 2 or 3 different landings and test them.
If one of the landings demonstrates the performance we need to make the campaign profitable or if a landing is at least very close to the objective, then individual and specific changes can be made on that specific landing to continue improving it.
The problem with trying 2 or 3 completely different landings is that, beyond speculation, it is VERY difficult to determine precisely which factor (or factors) made the winning landing perform better.
As a special bonus, I recommend that you download the following eBook about A / B Testing by Anne Holland that I translated into Spanish 🙂
TO DOWNLOAD "What's Worth Testing & What's Not"
Now it's your turn! Do you have an anecdote from one of your A / B Testing experiments? What is the most interesting experience you had with a campaign?
