Making the case for bite-sized PPC campaigns
When I received my PPC training, there was a prevailing belief that bigger was better. We would make gigantic keyword lists divided into zillions of ad groups, all in an effort to be hyper-targeted, hyper-specific, and hyper-relevant. The more important the client, the bigger the campaign we built them.
There is a clear downside to bigness, which is that it spreads your search traffic thin across too many variables. You end up diluting the very data you must accumulate in order to optimize your campaign.
What about the upside? The bigger-is-better philosophy hinges on the belief that you will get a steeper discount on each click the more precisely your keyword and ad text match the query. The assumed mechanism for that discount is the auction-time ad quality (wrongly called “quality score”), something that is poorly understood even by most experts.
In this post I will demonstrate why that assumption is wrong, and why most campaigns should be much, much smaller.
Google is smart and getting smarter
The whole idea of perfectly matching keywords to queries comes from an era when search engines weren’t very smart about interpreting user intent. Nowadays it often feels like Google is reading our minds.
The legitimate SEO world had to reinvent itself for semantic search when the old techniques didn’t work anymore. Yet the PPC world has largely carried on as if nothing has changed.
The truth is that the evolution of the Google algorithm affects both organic and paid search. Do you really think Google is so dumb that it needs fifty keyword combinations to understand that you are advertising plumbing services in Phoenix?
Or let me put it this way. If a human being asked what service you provide, would you list off a hundred synonyms and variations of the same thing?
Of course not. Nor would it make sense for Google to reward that kind of bizarre communication with cheaper clicks.
Reduce needless duplication
Google has stated repeatedly that their algorithms treat synonyms as identical where they signal the same user intent. So why build ad groups that Google sees as identical?
The same goes for singular vs plural keywords – when they signal the same intent, Google treats them the same. (This article gives some good examples of when the intent might differ.)
If we take Google at its word, then duplicating our keywords and ad groups with synonyms only dilutes the precious keyword-level data we are trying to accumulate. At a minimum, you should stop doing that.
Use location modifiers only when needed
Location qualifiers like “near me” or “in Phoenix” in user queries help Google understand user intent. Beyond that, it would be simple-minded for Google to care if a keyword shares those exact words.
That is, if the user query is “boat store in Phoenix,” Google is smart enough to know that there’s no meaningful difference between the keywords “boat store” and “boat store in Phoenix” if they are both location targeting Phoenix.
In fact, Google claims to treat geo-modifiers in your keywords the same as geo-targets. So if you’re already targeting an area, why needlessly duplicate your keywords?
Unless they’re a necessary part of a more complex location targeting strategy, you should ditch geo-modified ad groups and keywords.
Do not match stack your keywords
I was trained to duplicate my keyword lists in multiple match types. The theory was that more precise match types fetch lower CPCs because they tend to have higher CTRs.
In another post I argue that efforts to investigate that theory with data have tended to be deeply flawed. In a nutshell, empirical comparisons of CPCs across match types fail simply because the auctions are not the same. Broad match keyword X is eligible for far broader queries than exact match keyword X, so their metrics tell us nothing about how they compete head-to-head in the same auctions.
Moreover if you believe that Google is smart and getting smarter, there’s no reason to think it would reward strategies like match stacking. Google doesn’t make any more money if you have more keywords, nor do its users benefit when you have the same keywords in multiple match types.
Instead of thinking of match type as a way to get lower CPCs, think of it as a way to control the reach of your ads. By analogy, imagine a night club with two entrances and two bouncers. One bouncer is very strict, and the other will let anyone in. The permissive bouncer renders the strict bouncer useless, because anyone who wants in can simply go to the permissive bouncer.
Likewise your broadest match type will determine which queries your ads will show for. If you want more clicks you can broaden your reach, but be aware that you will face diminishing returns at the margin.
Choose your words wisely
The major premise of radical PPC minimalism is that adding keyword variations doesn’t increase your traffic, it just spreads it thin. Every keyword should have a clear justification for existence, or otherwise it just dilutes the very data we are trying to accumulate.
You should continue to organize your keywords into tightly-themed ad groups, making sure the root of each carefully chosen keyword appears in the ad text. Instead of trying to anticipate user queries, try to anticipate user intent. Think about what specific attributes might be important to a customer, and use those qualifiers to build meaningful keyword variations.
Be prepared for pushback
When your boss or (if you work with Ethical Digital and therefore have no boss) when your client sees the perfect diamond of a campaign you have built for them, they may be shocked by its diminutive size. They may get red in the face and wonder just what the hell they pay you for. They may pound their first and demand more keywords, more ad groups, more damn it more!
I’m not suggesting you dig in your heels and lose your job, but you can point them to this article and calmly explain the rationale to PPC minimalism. Remind them that a machine could add a million keyword variations, but only a master can hand-select exactly the right ones.
The juice is worth the squeeze
Ultimately your bite-sized campaign will be much easier to understand and manage, and since each individual keyword will accumulate data much faster, you will be able to optimize it far better.
I know it’s a radical departure from what most of us were taught, so rather than go full-minimalist right off the bat, you might try scaling down your campaigns one step at a time: first cut synonyms, then unnecessary singular/plural variations, then geo-modified ad groups, and then start adjusting the way you think about match type.
I promise you, your campaigns will perform better and your job will get easier the farther you go down the path of radical PPC minimalism.