I’ve been writing a lot lately about the growing phenomenon of “online-to-offline,” where consumers research a product on their mobile device or PC, and follow that up with an offline action, such as a call to set an appointment or a store visit to make a purchase. This type of spending will grow nearly 50% from 2012 to $1.7 trillion in 2017, according to an October Forrester Research Report.
If online-to-offline applies to you at all, probably the most effective thing to do is to measure offline action by each distinct keyword. Here’s why: Marketers often optimize online actions (like a form fill) by keyword because they are highly trackable, but then mistakenly estimate the impact of an offline action (like the value of a phone call) and apply that estimate to every keyword.
For example, let’s say that you’re a marketer for a cruise line, an industry in which customers research their options online, then make a reservation over the phone. Many campaigns will optimize around a click-to-call (literally, the number of phone calls), assuming the likelihood of a phone call converting into a reservation is equal or nearly equal for every keyword.
This is flawed logic, as any assumption of all keywords being equal will almost certainly lead to trouble. The fact is, the percentage of phone calls that have purchase intent varies wildly across keywords, so much so that optimizing an online action alone -- clicking to call -- as opposed to the offline action -- the purchase intent of that call -- will often lead to huge misallocations of budget.
To illustrate this, in the table below, I’ve compared the percentage of phone calls with true purchase intent for five brand keywords of a Fortune 100 advertiser that spends tens of millions on paid search annually.
How often a phone call is “good”
Brand KW #1: 13%
Brand KW #2 19%
Brand KW #3 21%
Brand KW #4 29%
Brand KW #5 42%
This data clearly shows huge variances based on the keyword involved, meaning that the search marketer would be making a huge mistake by treating each offline action as having the same likelihood to convert across all keywords.
A similar misallocation can occur when one assumes that same impact of an offline action from desktop or mobile. The table below compares rate of purchase intent for calls from desktop and mobile for the exact same set of keywords.
Desktop Calls w/ Purchase
Intent Mobile Calls w/ PI
Advertiser #1 - 26% 16%
Advertiser #2 - 12% 42%
For advertiser #1, a communications company, calls from the desktop convert significantly better than calls from mobile for the same set of keywords. For advertiser #2, a home services brand, the reverse is true: calls from mobile perform significantly better than calls from desktop for the same set of keywords.Numbers like these provide early proof that in the world of online-to-offline, each keyword should be treated distinctly -- and that diligently measuring online-to-offline results from search can provide a lot of low-hanging fruit for optimization and improved return on ad spend in 2015.