Of the 21 Secret Truths in the book, #4 was in some ways the most difficult to write. It’s one of the most abstract ideas in the book, perhaps the least intuitive, and it really needed at least a couple of full-blown examples.
All of which made it tough just due to the space constraints available. The book was designed to be quick and easy to read, and every entry was allocated exactly one page. But on this one I really appreciate the chance to offer extended remarks and comments.
The idea is simple, if hard to compress into a single sentence: Campaigns don’t do anything to your account performance, they’re constructs that should make your reports more informative and actionable. As such, using them simply as the top level of a hierarchical logical organization of your keywords is a waste.
What Campaigns Do and Should Do
In our experience, campaigns are most typically used to define and create a categorical breakdown of an account. A clothing retailer would by default have a shoe campaign and a sox campaign and a hats campaign. Inside of those would be the associated ad groups.
The result of this is that when you look at campaign reports and performance – which you do all the time because both AdWords and 3rd party tools like ClickEquations naturally present the campaign level data to you quite prominently – you see the summary performance (impressions, clicks, revenue, costs) for the campaigns based on those groups.
Here’s the problem: Seeing the results rolled up based on those categorizations isn’t very useful.
Sure it seems nice to know that shoes has a 4% CTR and a 250% ROI while hats has a 5% CTR but only a 150% ROI. But is it really useful?
The problem is averages. What you see in these rolled up results, perfectly reasonably, are averages. On average in the shoes campaign the CTR was 4%. And as we’ll discuss in more detail in a later post, averages are the enemy of accuracy. They mask facts and trends by their very nature.
Average can be put to great use – they’re statistically useful. But they can be inappropriate too.
What’s hidden in typical categorical organization is the clarity you can get if you further break down campaigns based on other, additional, distinctions within your campaigns. Basically you want to think about the different types and classes of ad groups the campaigns contain, aspects that would cause dramatically different performance, and collect those types of ad groups into campaigns based on those similarities.
Breaking Campaigns Down
There are several types of ad groups that you might want to segregate. Ad groups that contain brand keywords are obvious. Brand keywords get vastly higher CTRs, better conversion rates, and often lower CPCs. If you have brand keywords mixed in your general campaigns, they’ll really distort the average numbers reported.
Often within different target product or offer segments in your businss you’ll have keywords that carry different business intents. You may have some keywords that bring in a lot of new customers, generate high traffic volumes, but aren’t very profitable. Call them loss leaders, or new client introducers, or just keywords aimed at revenue more than profit.
On the other hand, you likely (hopefully) have some keywords (well organized into tight ad groups!) that just kill it in terms of pure good old profit. They have your best conversion rates, highest average order values, and for them you manage very tighly to maximize these already good returns.
If stuffed within your shoes campaign are some ad groups that are high volume but marginally profitable, and others that are super profitable, doesn’t that doom the rolled up campaign results to be rather meaningless? What are they going to tell you?
By contrast, suppose you take very small number of mega-profitable ad groups out of the shoe campaign, and make a new campaign called ‘shoes-high-margin’. Now every day/week/month, you can look at those campaign stats and quickly get an accurate idea of if that profit gravy train is on track. If there is a dip, you’ll see it quickly. If there is a surge, you’ll know that too and can respond with more budget or perhaps even more keywords.
Lousy performers need the same treatment. We all have keywords (and perhaps ad groups) that just aren’t doing well. Maybe we should kill them but just don’t have the heart. Maybe we’re working really hard to test better ad copy and tweak negatives and match types. In any case, for now they’re losers.
Mixed into our everyday campaigns, the losers hide in the shadows. We don’t clearly see how much they’re really costing, or how far below the averages they are. Often they live on for months and years. Drag them into their own campign, get forced to stare every month at $29,000 spend and $1213 revenue, and your motivation and decisions just might change.
Plus, the reporting on your core campaigns, minus these misfits, is much more accurate too.
Looking For Wow!
Hopefully this clarifies the point. If campaigns are simply logical categories they’re data is of limited use. If they’re grouped logically and by performance or at least goal then the numbers they produce are meaningful.
Here’s the real goal: You want to be able to see a number in your campaign report and say either WOW or OH SHIT. There should be numbers in those reports that have expected ranges and reasons behind them, and if they change you should be able to know that it’s a big deal.
If they’re all huge roll-up averages that jump around, or that stay constant because even huge swings within them are masked by other shifts elsewhere, there will never be and Wow or Oh Shit moments based on campaign reports.
And beyond saving you that little drama, it means that important things are happening in your campaigns, and you’re missing them.
We don’t want that.
There are two related topics: How to best actually reoganize campaigns, and the impact of Impression Share on the campaign organization decision. I’ll tackle the first one in a follow up post in the next few days, and talk more about Impression Share and in terms of campaign organization when that topic comes up in the natural sequence.
What Do You Think?
This blog post is part of a series extending and amplifying the ideas in our free ebook ’21 Secret Truths of High-Resolution PPC’.
What they’re saying on twitter: “Very, Very, Very nice e-book from @clickequations called ‘21 secrets to PPC’. Easy to read, and full of good and funny stuff! – @Eloi_Casali”