UPDATE: Nearly two in five CMOs say their biggest challenge is “integrating and tracking multiple channels,” according to a recent survey by Aprimo and Argyle Executive Forum. In light of that finding, I thought I’d reprise a column that I wrote last year and use it to discuss correlating relative ROI (define) across channels like TV, print, radio, and online banner ads.
Lift in search volume for particular terms – like non-generic brand names – is useful because it indicates that the target customers not only saw the ads, but also remembered them and then took action. It is a better metric than those that only estimate whether ads were even seen or heard – i.e., reach and frequency-based metrics. This even works well for online banner or branding campaigns because it is not the click-through from those ads that we are concerned with, it is the lift in search volume that indicates interest and intent. The habit of search by modern users is now commonplace; as they all go online to look for additional information before making a purchase decision, we now have a universally applicable metric to correlate the effectiveness (and thus the relative ROI) of advertising across channels.
What about absolute ROI? We’ve got that covered too. Beyond lift in search volume, with proper analytics on online destinations, advertisers can track where customers came from, calculate how many took desirable actions like completing a purchase or printing a coupon (for items purchased offline). With these metrics, advertisers can correlate with other known data such as conversion rates of coupons, etc., to derive real-world ROI, scientifically. Obviously there are nuances and caveats, but if applied carefully these methods can help CMOs overcome what keeps them up at night.
So let’s return to the discussion I started last year in the column that was originally published March 12, 2009.
What is the return on investment of search, social marketing, or digital marketing or advertising in general? To me, ROI had always been calculated from sales. And I was curious why advertisers and marketers were not constantly asking for ROI for television, print, and radio campaigns while they kept harping on it with digital and social marketing. Then, it occurred to me that advertisers and marketers could have very different definitions for the term ROI. Perhaps there is indeed a way to reach common ground on the definition of ROI and use it to compare campaigns that run in different media or channels.
The ROI of Traditional Media? You Kidding Me?
To some, return on investment meant the reach and frequency they achieved (e.g. on television); to others it was the number of impressions of display ads that were delivered; to others it meant the number of clicks through to a Web site, unique visitors per month, time spent on-site, or even number of fans on their Facebook page. And they would perceive that they got a better ROI if subsequent campaigns achieved greater reach and frequency, impressions, unique visitors, etc. or they achieved the same level at lower cost.
More often than not, these parameters cannot be directly related to sales. For example, a television ad reaches incredible numbers of people — perhaps even the right people. And, studies will even show greater brand recall or affinity. But no one knows for sure whether the customer recalled the message sometime in the future when they were actually about to buy whatever was being advertised. The standing joke in the advertising industry, attributed to department store merchant John Wanamaker decades ago, goes, “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”
Well, that’s just not good enough. And to imply that “half” the advertising does work is probably 50x too generous. There are some sandpaper approximations (i.e. very rough) that use surveys (“did you end up buying the product”) or panels (“what percent of the panel ended up buying the product”) to attempt to correlate the above metrics to sales. For example ComScore requires a brand to have at least 80 million display-ad impressions before it can even give an approximation of sales using their panels. Why? Because the intersection of people on their panel with people who may have seen a display ad is so minuscule, they can’t find enough people to ask “did you buy the product?”
These rough approximations also fail to take into account real life. The difference of a few people saying they did end up buying (versus not buying) out of the small number of people on the panel, extrapolated to tens or hundreds of millions of impressions will lead to massively different answers in the approximation of ROI. These methods do not account for outside forces like:
- I saw the ad, I liked the ad, I recalled the ad — but I just bought a car and won’t buy another one no matter how much I liked the ad.
- I saw the ad, I remembered it, I went online to do more research — but I could not find the answer to how this quad-core microprocessor was different than another quad-core microprocessor that was half the price so I just bought the cheaper one instead
- Or, I saw the ad and recalled it, but I don’t need a new insurance policy until next year when my current one expires.
Relating Awareness Advertising to Online Consumer Actions
Many advertisers still use the approximations of reach and frequency as a measure of success or ROI for campaigns. Others are starting to use more precise measurable online metrics such as banner ad impressions, unique visitors to a site, page views, or time-on-site as campaign goals or key performance indicators, or KPIs (define). But these metrics are still not great indicators of program success, especially if ROI is calculated from “sales.”
In the column, “Social Intensity: A New Measure for Campaign Success?” we propose search volume or social intensity as better indicators of not only recall, but also interest and intent — e.g., people search heavily just prior to a purchase and they ask friends for input about the impending purchase. But how do we know who is searching and when? We don’t (ahead of time). But the point is, when they do search, that’s the best time to provide consumers with the information they need to make an informed purchase decision and help move them expeditiously toward the purchase. This is why search advertising is rising and display advertising is waning — search simply maps better to modern consumers’ well-demonstrated behaviors. This even applies to low-consideration, low-cost, and impulse products — modern consumers will even search for nutritional information for canned soup or pouch tuna.
In fact, there is finally evidence of TV advertising’s effectiveness. The Super Bowl ads of Denny’s, GoDaddy, Vizio, and Hyundai drove detectable lift in search volume around those terms for the day of the Super Bowl and the following two days. There is also lift in search volume upon launch of a new product, especially noticeable when a made-up word is used — e.g. Subway “Footlongs.” There was no search volume on “footlongs” before the ad campaign and there is a spike in search volume when the campaign is running. In each example, the viewer not only saw the commercial, but they recalled it, and they made the effort to go perform a search online. People who are not considering a purchase are unlikely to have taken the trouble to do so.
Using Search to Relate Awareness Advertising to ROI
So, once you can correlate awareness advertising to an indication of interest and intent — search volume during the time the ads were being aired — then you can correlate the lift in search volume to the lift in a site’s unique users, again during the same time period.
Assume the site is well SEO-ed (search engine optimized) and assume that certain search results map to specific pages of the site that are specifically useful to the person searching. Then, there’s a way to correlate that visitor or group of visitors to a sale (e.g. e-commerce on the site) or a proxy for offline sales like printing a coupon, whose subsequent redemption rate is well known from the advertisers’ own historical data.
With this correlation, you have a pathway that leads to sales that are attributable to specific campaigns, and hence a way to calculate ROI. This pathway may even corroborate other methodologies that used rough approximations to calculate ROI, and thus lead to greater confidence in the resulting ROI number for any campaign.
Tips for Running Awareness Campaigns to Calculate ROI
- Ensure that online destinations are set up prior to launching paid awareness campaigns. So when people go online to search and do research after being inspired by the awareness campaign, they will find you and the specific information they are looking for.
- Use made-up words when possible, so lift in search is detectable. Brand names that are also generic words like axe, tag, open, etc. will not be detectable about the noise of unrelated searches containing the same word.
- Search-engine optimize your site and online assets, so specific information can be found and is mapped specifically to the “missing link” the potential customer is trying to solve.
- Have detailed analytics in place to measure the “conversion funnel” that can end in a sale or another action and is a known proxy for sales (e.g. advertiser knows X percent of coupons printed online get redeemed with a purchase offline.)