Creative Testing – a full time affair

Okay, we looked at setting up the campaign, ad group, keywords, as well as the metrics, tracking and reporting. One of the often over looked opportunities with Google and now with Yahoo Panama is the ability to test creative. However, it is not just a matter of putting out a bunch of copy executions and seeing what happens with the CTR. You need to think these things out in advanced and develop a process or system to test these affectively relative to your ultimate online goals.

Whats this about
Before I go further, I want to be clear. This is not “how to write good copy.” This posting is about how to set your program up to find the optimal keyword, copy and experience combination. Too often, I hear about copy in terms of what drives the best CTR. When, in fact there are instances where you want users to self select out and effectively lower your CTR. Why? Because this is about the ROI…what happens post click. Getting to the best copy in terms of ROI / ROAS is what this post is about.

Identify the Consumers' Drivers and Your Control
As mentioned in previous post, associating subject matter / words in the ad copy with the landing page is very important to improving the quality score thereby lowering your CPC and improving your position. However, there is a lot more to it than that. If the whole theory of the quality score plays out, then a higher quality score should also improve your conversions. Whether this is true or not can only be tested by you (I have seen and read of many situations where the quality score was low, but conversions were high).

To assess your ability to manage quality score along with optimizing conversions, begin by identifying those things related to your site over which you have control. Ask:

1)    Can you drive to any page you want and still have the tracking you need?
2)    Can you control the content of any particular page?
3)    Can you add pages?
4)    Can you control offers or pricing?

Then ask about what drives the desired action. What are users looking for that will drive the actions you want (buying, requesting more info, signing up for newsletters, etc)?
1)    price
2)    selection / availability
3)    easy of process
4)    quality
5)    actionable information
6)    brands
7)    attributes
8)    Other?

You probably know what people talk about regarding your product or service and this is a good starting point. However, if you have not already tested other messaging subjects, don’t rule them out right off the bat. In other words, challenge what you “know.”

Okay, now the interesting part. Take a look at the list of what drives consumers and take a look at your site. This is where your level of control comes in. Worst case scenario is where you know that nothing on your site speaks directly to the consumers’ drivers and you can’t change that. Usually, it is slightly better than that. You have some sub-pages that have relevant subject matter, but over which you have no direct control. If you’re one of the fortunate few, you have actual content control. Where ever you are on the spectrum, be sure you understand it. It is important when mapping out the next steps.

Where to send the users
So, we’ve covered a fair bit, and I have not written one word about actual ad copy. Before you start writing ad copy, you must have a clear picture of your users’ drivers and your own control. You want to ensure that the users’ expectations pre-click match their experience post click. To do this, map out the drivers of behavior to the most appropriate place on your site.



If the productSelection.html page is the only one to which you can drive traffic, this may be good for some things but poor for others. This is where your control level comes in.

What you want to do is be able to have specific pages developed, either dynamically based on parameters, or with static HTML that you have the option to change quickly. So your matrix can look something like this:



In this scenario, the price driven users land on pages with well targeted subject matter geared toward promotions, rebates / price, quality, and others attributes. Its obvious; if there is a key driver that gets them to click on the ad, then why take them to a page that has nothing to do with it? Drive them to a well targeted page.

Once you’ve done this, ask yourself yet another question: “does the section of the site your users end up on really speak to the issue / subject you believe drives their behavior?” The answer ought to be yes. If not, can you exercise some control? For instance, with pricing driven shoppers, some may be moved by straight sales, others may be moved by “package” deals. So between, “25% off the pants”, or “buy these pants and get 50% of selected shirts,” which drives more revenue per ad dollar?

Copy Content Sources
Now, you have your “drivers,” your target pages and your tracking (or you will.) I don’t know what the desired event is for your site, so obviously I can’t give actual copy. But, what I can do is help you wrap some structure around your ad copy practices. For the copy itself, look to several sources:

1)    The competition. Not to mimic them, but to see about points of differentiation. There may not be any (hard to tell a boss that, so come up with something), but it is good to stay aware.
2)    The web site itself. What is on the page? Look at logs…what text or images drove clicks to your target page from within the site?
3)    Off-line, POP, direct mail, print ads, etc. for you and the competition.
4)    Marketing and advertising research / customer feedback.
 

You get the idea…you’re not isolated. There are many sources of information related to copy development. Now for each of the drivers, develop at least a few copy options.

Setting up Your Test Matrix
Then create a matrix that will allow you to start tracking results.




The above matrix is very basic. For each driver, you may have multiple test landing pages, or you may send the same driver to multiple landing pages, or have different metrics. Keep in mind that the value may be realized from an immediate sale, or from a sign up for a newsletter leading to a later sale. If your site has both, then the metrics are different for each. The point is, look at this as a starting point and mold it to your own situation.

Beyond that you should take this down to the keyword level. This is where the nuances play out that can lead to the epiphanies. You may find that the same drivers create great sales for one keyword and not well for another in the same ad copy / landing page combination. The poor performing keyword may however do well with different ad copy going to the same page, or same ad copy to a different page or…. The list can goes on. That is why you need the matrix and keyword level tracking.

Keep Testing
For the copy itself, keep testing. Consumers change, seasonality can move things on you, competitors may influence the market behavior. Always be willing to test what you think you know. Assume things have changed.