3 Ways to Use Data to Advance Your Search Performance
No matter how much we may gripe about the engines’ "black box" algorithms, we can't argue with performance and a lot of deep data that we can get our hands on.
For some, that might be a curse. Like a paella-laden Estelle Costanza might say, "what are we supposed to do with all this data?"
Use it. And keep using it. Then find more data, and use that too.
Test, implement, analyze.
Lather. Rinse. Repeat.
1. Expand Your Keyword Lists
Building out and refining your keyword lists is a good example. Everyone knows the value of the long tail in helping lower CPCs. And yet we all know the tendency to throw everything in there and come up with a list that's unwieldy and impossible to manage.
When it comes to finding the right keywords for your account, there are lots of tools available, whether free options from Google, a third-party system, or an agency's own proprietary tool. But there's a lot of value in pouring through search query reports to find long tail and negative recommendations.
By matching the query to the credited keyword and then analyzing the return on that query, you can identify where to find lower CPCs or reduce waste by adding negatives for non-converters.
A more refined approach is to leverage the conversion rate, cost and revenue to come up with a projected contribution margin for typed keywords. Then match those with what's currently in your account. Missing keywords with high CMs go into your long tail; negative CMs go into your negatives.
For example, if you have a keyword "Acme brown widgets" and find that Google is matching up the query "extra large Acme brown widgets" to it, look at the performance of that keyword for each time it's matched to that query. Subtract cost from revenue to find the contribution margin (or divide if you prefer to organize by ROAS) telling you how profitable that query is.
If it's high, add the longer tail "extra large Acme brown widgets" to your keyword lists and the ensuing lower CPCs will bring even greater margin. If it's lower, "extra large" may be a good negative keyword for this campaign.
Just be sure to run a similar report after these new keywords have garnered traffic and refine as necessary.
2. Get Your Accounts (More) in Order
Another consideration is simply how to organize your data. You've committed to the campaign structure that makes the most sense, but it's never the only way you need it broken out.
You can apply lots of filters, merge multiple reports or write custom scripts. But those can be inefficient or simply unavailable. If your platform allows you to leverage virtual group structures, you can organize your account multiple ways simultaneously.
A retailer, for example, may have a brand/non-brand campaign structure, with individual product categories at the group level. So if Acme produces five different widgets for them, those could be spread across five different groups (10, if you consider some could be brand and some non-brand and yet more when you factor in multiple search engines).
By setting up an "Acme" virtual group, you can bundle those related but separated groups together in an "alternate" structure for more efficient campaign management and reporting. You have easy access to all of the related creative and keywords and can have Acme virtual group reports automated to be sent to product managers and brand marketing teams.
A travel advertiser may find the same thing when trying to group together different geos, lines or price points.
With virtual groups, it's easy to analyze performance across different cross-sections of your account or pull relevant reports for various stakeholders.
3. Utilize What's in Your Cart
Retailers are used to tracking the value of every conversion. And while many may go so far as to track the individual items that were in the cart, surprisingly few do anything of real value with that information.
The purchase amount helps you optimize bids. But analyzing the items in each sale and matching those back to the keywords, ads and landing pages that led to the conversion can help inform your strategy for creating and revising these assets.
If a keyword or ad has a nice ROAS, you may be doing well from a bid optimization standpoint. But what if you’re still not selling enough of your higher-margin widgets? This can drive creative and landing page testing to increase the return of a group you previously thought was doing just fine.
Taking this a step further, you can also leverage onsite data such as the buying propensity and specific product interests of users on your site. This data can be used to help automate some of the testing and increase conversion rates, as well as inform you about the individual product interests of your customers by keyword or ad. A person might search for "fancy widgets," click on your fancy widgets ad and then browse your site for awhile, finally seeming most interested in a line of higher end fancy diamond widgets.
Armed with this data, future messaging to that visitor (remarketing, website content, etc) could more appropriately be targeted to their true product interests. We consistently see that more relevant messaging both on and off a marketer's site results in higher conversion rates, so finding ways to utilize this data can be beneficial.
There's a lot of data out there – what do you do with it all? Use it. Luxuriate in it. And then use it some more.