The higher up the PPC game you look, the less difference you’ll find between practitioners.
Some things almost always work
Some things almost never do
Some features are rarely useful, but might be just perfect for a particular type of account…
Some metrics matter a lot less than they seem
The more experience we gain, the more we understand – and converge around – what falls under each of these categories.
But true to the narcissism of small differences when we have a chance to disagree, we take it with fervour…
As you probably know, SKAGs are a topic dripping in controversy.
It’s been that way to some extent since SKAGs became ‘a thing’ in the early 2010s… but the controversy has grown and persisted into 2021 and – the signs suggest – beyond.
With so much being written about SKAGs and how great/limited they are, an opinion piece on SKAGs seems a little late to the party, and yet
a) Posts like this one (from my Facebook group) still show up with surprising frequency:
b) They invariable attract a lot of attention when they do
c) There are still plenty of voices on both sides of the argument
So there still seems to be an appetite and – I hope – some use for a general treatment of the SKAG issue.
First, let’s quickly run through what a SKAG is…
SKAGs are single-keyword ad groups.
Instead of aggregating keywords around a particular theme or intent, we place each keyword into its own ad group, for the purpose of super-granular control and refinement of the keyword > ad relationship.
These ad groups are then often kept discrete by the use of ‘internal negatives’ to ward off search term overlap, and ensure that each keyword (and ad group) covers only its intended territory. More on this later…
When SKAGs hit the PPC scene, they were widely seen as a sort of wonderful ‘hack’.
They were never what Google intended or advised (the idea of the ad ‘group’ was always to aggregate keywords with sufficiently similar character to use the same ads)
But it did take the principle of maximum relevance to its logical conclusion, and it became an extremely popular and well-respected structural framework.
Then things changed.
The phrase ‘SKAGs are dead’ is one you won’t have to wait long to see if you hang around PPC groups and forums.
In my view the SKAG method has always been questionable (although for a while skeptics like me were swimming against the tide…) – but this wider exodus has been prompted by the latest few rounds of changes to how match types work.
Specifically, keywords of all match types have incrementally been allowed to catch a wider range of search terms, compromising advertisers’ ability to herd those search terms into precise segments with keywords.
(and in the end, search terms are the active ingredients of keywords. As search terms become less closely linked to keywords, keywords become less well-defined and effective as agents in our accounts… asking them to act alone then becomes a less powerful option.)
This blurring of the scope of / boundaries between keywords has undoubtedly made SKAGs – at the very least – much harder to manage. But it’s just one of many factors to consider in weighing them up.
There are real advantages to SKAGs, and there are (and always have been) major disadvantages too.
Let’s look through them:
The logic of the SKAG is solid. It is based on the undeniable benefits of messaging congruence.
When the search term leads to a highly-relevant ad, and the click leads to a highly-relevant page, everyone is happy. (‘everyone’ in this case being the user (CTR, UX and CVR), Google (QS) and you (£)).
SKAGs are clearly a way to achieve keyword-ad-text harmony throughout your account.
And SKAGs can – and often do – perform well in practice.
If done right – with internal negatives (though this is an increasingly tail-chasing exercise thanks to the erosion of match type boundaries) – SKAGs can give you real search term clarity. That is, they will let you see clearly which search terms are acting on which keywords (and therefore in this case ad groups and ads…)
This in turn gives you more discrete levers to pull, while knowing precisely what input (search term) you are working with.
A tight keyword-ad relationship is one of the central features of account success. And SKAGs are a way to achieve it.
Credit where it’s due: This is no small advantage.
There are three broad reasons why – despite the advantages – SKAGs are less than ideal as an account structure.
I called it an advantage that SKAGs give you more discrete levers to pull.
That is… they give you more entities to work with across the same level of the account hierarchy.
What they take in return is an entire layer of that hierarchy.
When keywords and ad groups are effectively the same thing, we lose an increment on our ‘zoom’ function… we can no longer clearly see or simply make use of performance patterns at any level between the individual keyword and the campaign.
What we do every day as Google Ads account managers is look for performance patterns and act on them (or not) in the best way we can. A structure that obscures those patterns by stretching its contents into a string of atoms poses a real problem for optimisation.
In practice this atomisation also means that close, clear keyword-ad matching takes a shed load of work to maintain as an account changes, particularly if you’re trying to scale it.
A lot of work or – more likely – a structure that breaks…
Apart from presenting a less navigable structure (and the real-world downside of an account that’s uncomfortable to manage is not to be underestimated…) the data fragmentation also risks a more direct disadvantage to performance.
Just as you lose the analytical advantages of aggregation yourself, so does Google, and its various algorithms.
For example, splitting activity into too many individual ad groups means that individual ads often won’t gather sufficient search volume and data to build any significant history (either for actionable data patterns that you can use, or for preferential treatment by Google’s algorithms).
By aggregating keywords within ad groups, you can allow your ads to gather data faster and bring those insights into play.
And when it comes to smart bidding – while Google is perfectly capable of aggregating learnings across ad groups and campaigns (with or without portfolios) – the unit by which Google is now recommending we maintain minimum volume levels is the ad group.
From one of Google’s recent presentations:
Where SKAGs have a place
None of this is to say that keywords should never be put into their own ad group.
Sometimes there’s a good reason to focus on one particular keyword for specified ad text, dedicated targeting options or bid adjustments… In this case, there’s no advantage in adding more keywords to keep it company.
Occasionally high-priority keywords can even warrant their own campaign, to differentiate budget or location targeting.
Tightly-themed ad groups have always been a genuine best practice, and it pays to place more importance on tighter divisions with higher-priority keywords. For the highest-priority, these can reasonably become single-keyword ad groups.
But SKAGs are not an ideal way to structure a whole account. They lead to too much fragmentation of data, and a sprawling structure that’s hard to navigate and manage.
The principle behind SKAGs has also become less appropriate as the role and nature of the keyword has shifted over the past decade.
The idea of atomising your account at least made sense when keywords could be thought of as discrete units with reasonably clear boundaries.
Each time Google rolls out a further change to match types – always at the expense of the clarity and precision of the keyword <-> search term relationship – another swathe of SKAGers loses faith.
Google is pushing us in general to reduce the granularity of our account structures… to aggregate, cast a wide net (upgrade to broad match, anyone…?)
and put faith in the algorithms’ ability to make use of patterns that we wouldn’t even have the available data to find if we tried…
But there are also issues with this ‘Hagakure’ approach, which sits on the other end of the granularity spectrum.
So the search for the best middle ground is ongoing, and the factors involved will change as the platform continues to shift. They will vary between accounts too.
One thing that should be uncontroversial by now though, is that we won’t find a single optimum account structure at the extremes.
If you want to go deeper into account structures…
- What factors to consider when dividing your account into campaigns and ad groups
- how to present an account structure to clients
- more on SKAGs
- (and what the Roman legion has to teach us about account structure!)
You’ll find all this in section 6 of the paid search optimisation course in Google Ads Level Up.