Head vs. Tail – How To Shift Your Priorities For Each Type Of Account
“There aren’t enough hours in a day” is an old – but true – refrain about the nature of an SEM’s job. The list of things we could be doing in any given day to optimize accounts is a daunting one. Agreed? Fantastic. As long as we’re on the same page, and already slinging the […]
“There aren’t enough hours in a day” is an old – but true – refrain about the nature of an SEM’s job. The list of things we could be doing in any given day to optimize accounts is a daunting one. Agreed? Fantastic.
As long as we’re on the same page, and already slinging the clichés, this post will help you “work smarter, not harder” by helping you identify where your accounts lie on the head-tail spectrum – and how to use that knowledge to prioritize tasks like ad testing, keyword optimization, negative keywords, and more.
First: Identify If Your Accounts Skew Towards ‘Head’ Or ‘Tail’
Head accounts are those where non-branded KPIs (primarily conversion volume) are condensed within a relatively small query pool. These are the accounts that follow the Pareto principal – that is, 80% of your output (sales, revenue, etc.) is generated from 20% of your account (in this case, queries).
Even this, however, tends to overstate the situation, as 20% of queries by count is still extremely large for a true ‘head’ account. In fact, as few as 1 or 2% of unique queries and/or keywords can generate up to 80% of sales in true head accounts!
In these accounts, each individual top non-branded query can be responsible for 20% or more of non-branded conversions.
Tail accounts are those where no individual non-branded query makes up more than about 2% of total non-branded query conversions. These are accounts where your top non-branded query can fall off the map completely, and odds are you wouldn’t notice it.
In these accounts, you’ll find thousands of unique converting non-branded queries, and the list of converting queries would be extremely large.
In the graph below, I’ve sorted the top 50 non-branded queries by conversion volume from one head account and one tail account. Then, I chose to graph the cumulative percentage of conversions brought in by these query sets relative to the entire set of non-branded queries.
To get an even better sense of how dramatic the difference between Head and Tail account is, in the head account the top 5 non-brand queries make up 44.25% of non-branded conversions, while the tail’s top 5 make up just 5.01%.
If I were to continue the graph of the ‘tail’ account, it would take roughly 7,000 queries before we approach the percentage of conversions brought in by the ‘head’ account’s top 50.
Based on these differences, it starts becoming apparent that a one-size-fits-all approach to paid search isn’t optimal. Throw in things like management of branded queries, as well as the entire Google Display Network, and it’s impossible to come up with a list of ranked prioritization across the breadth of accounts in different verticals.
Going off of these benchmarks, you should be able to identify if your account skews towards a head account or towards a tail account. While the accounts that I’ve chosen are close to upper and lower bounds, I’d actually expect that most well-built accounts look similar to my examples.
That is, few accounts would fall directly between a head or tail account. Even so, accounts will run the gamut, so, instead of classifying an account as ‘head’ or ‘tail,’ it’s typically more appropriate to consider the degree to which an account is head or tail.
So what use do we make of the classification?
Well, while each and every optimization effort is important for accounts, the prioritization of them differs depending on whether or not the account in question is ‘head’ or ‘tail.’
Note: Remember that an account can be classified as ‘head’ or ‘tail’ only after reasonable keyword expansion has taken place (to avoid false ‘head’ accounts, which don’t have enough keywords built out to capture long-tail queries), and the account has consistent distributions over an adequately long time frame.
Now, on to the tasks – and how to prioritize them based on head/tail categorization.
Typically, ad testing gets flagged as the most important aspect of PPC. That makes sense, since it’s all that users see on the SERPs. However, the impact of ad testing much more easily felt on head accounts.
On a head account, an ad test that results in 20% improvement in CVI (conversions per impression) to the top 50 queries nets out 8.8% non-brand improvement – not bad for a day’s work.
On the other hand, the same 20% improvement on the top 50 queries of the tail nets out only about 1%. Couple that with the fact that ad testing is slower in tail accounts, since the clicks accrue much slower for the same quantity of queries, and ad testing is a much more important task for head accounts.
Now, if a PPC manager could find a test that works across keywords and product lines in a ‘tail account,’ then the successful test can touch a higher percentage of the non-brand portfolio than just the top 50 queries.
In tail accounts, however, you are typically dealing with many different competitors, auction conditions, margin goals, etc., so it’s rather unlikely that an account-wide roll-out of a successful ad test in a tail account will have the same results on the rest of the non-brand portfolio.
Additionally, the data on ad copy tests in head accounts is much cleaner. In fact, you can do ad copy tests for each head query and still end up with enough data to make a statistically significant choice!
The skinny: critical for head accounts; important for tail accounts.
Keyword expansion is primarily a tool for capturing new queries. Since there is a much larger pool of unique queries in tail accounts, it’s much more important to focus on expansion here.
This idea extends to match types as well. Since head accounts have such a high concentration of conversions occurring in a small pool of queries, it’s feasible to run less aggressive match types, such as exact and phrase match. Conversely, since there are always more conversions to be found in ‘tail’ accounts, it’s advisable to run broad and modified broad match with regularity.
If asked the question ‘What are you failing to capture?’ in a head account, it’s unlikely there would be much of anything. Even if only half of the tail were being captured in a ‘head’ account, that’d only leave a small percent of new conversions to be found.
In a tail account, there are tons and tons of unique converting queries. As such, it’s feasible that there are new veins of traffic to be found, and that if enough of them are found, it could result in a very tangible traffic increase. So, at the minimum, tail accounts have the potential for growth via keyword expansion, whereas head accounts have little potential for growth.
Semantically, it’s also much more plausible that the account manager simply ‘forgot’ a keyword in a tail account, as opposed to a head account, where there is likely only on product or service being offered.
The skinny: critical for tail accounts; not very important for head accounts.
Much like positive keywords, negative keywords are a more important task for tail accounts. Since more queries are being captured in tail accounts, the likelihood of more poor queries being captured also exists. While this speaks to the quantity of negatives, high-volume negatives will almost certainly exist in both types of accounts.
Finding the high-volume negatives is a rather finite task, so early on each account will have equal dependence on negatives. Going forward – that is, as a recurring task – tail accounts will consistently find new low- to mid-volume negatives, whereas head accounts may find nothing at all over time.
For either account, however, it’s important to have as much of an impact on the account as possible. That is, instead of negging out a 1-impression 1-click query, identify the token that is irrelevant, and apply it across the account (as a general rule).
The skinny: critical for tail accounts as both initial and ongoing task; critical for head accounts as an initial task but not very important as a recurring task.
Bidding as a whole is very important for head accounts and tail accounts. However, the integrity of bids on head accounts is substantially higher. That is, it’s easier to determine a ‘true’ bid for a query with high volume than to determine a bid for a cluster of queries in the form of an ad group-level bid or broad match keyword.
Additionally, since there are fewer optimization levers for ‘head’ accounts, the relative importance may well be higher there, but the absolute influence will be similar for both accounts.
The skinny: critical for head accounts; important for tail accounts.
Making Heads Or Tails Of It All
Head accounts and tail accounts are very different. As a general rule, when considering an optimization effort, I think ‘What percentage of my conversions is this action likely to influence, by how much, and in which direction?’ This type of sanity check really keeps good PPCers from getting into a one-size-fits-all mentality.
Now, after reading this, you might wonder “Would I rather work on a head account or a tail account?”
If you’re a creative person who writes killer copy and obsesses over landing-page mirroring and minor product tweaks, go for the head account. Make sure you’ve got a keen attention to detail, in the form of precise ad copy tests and minute bid changes.
Also, don’t be surprised if one day you wake up and some crazy competitor increased the auction prices by 30% on your top 10 queries! Your top 10 are also most likely their top 10, there are no surprises or ‘hidden gems’ in head accounts. Perfectionists prefer head accounts.
If you’re a grinder who doesn’t mind bulk keyword builds and you aren’t paralyzed into always waiting for statistically significant data to make changes, go for a tail account. Tail accounts can always find an extra point or two, and the level of ‘depth’ required on an individual query is less important. Go for a tail account if you’re always starting new tasks before perfecting the last one.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.