Sep. 30th, 2006

Calculating online and offline ROI: some tips from the Mortar.


Above: Subservient Chicken attracted 14 million visitors who averaged of six to seven minutes on the site, and
it became a pop culture favorite. The chicken even served as host for
Fox’s Sunday night lineup. But asks Adweek: "did the campaign actually
drive customers into [Burger King] stores to buy the sandwich? About a month after
the TenderCrisp sandwich debuted, BK reported that sales had steadily
increased an average of 9 percent a week. Since then, Geis says the
company has seen "double-digit" growth of awareness of the TenderCrisp
Chicken Sandwich and "significantly increased" chicken sandwich sales.
And the TenderCrisp does sell better than the Original Chicken Sandwich." Keller says he knows only anecdotal results, but to
him, they lean toward the positive. "Foot traffic is my biggest
concern," he says. "I want people going to BK, talking about BK. I got
a call from a friend of my wife. She said, ‘I was in Burger King and I
don’t know why.’ That to me is an effect of advertising."

A journalist called today looking for input on an upcoming feature about measuring the results of online and offline marketing.

Consider Jane. Awoken by traffic reports sponsored by Bob’s Trouble Free Mufflers she scans her morning newspaper over coffee — stopping for a few seconds to scan an article about car repair shops that rip-off women. Bob’s Muffler’s CEO Bob Wilkinson is quoted extensively. Maybe she catches Bob’s commercial on TV before heading out the door. On the way to work she passes numerous billboards, one of which reminds her that her nearest Bob’s is just around the corner. 

At the office she checks her personal email, surfing a couple of sites before settling down to work. If we’re lucky she follows a link to Bob’s site. But then she’s distracted by an email and inadvertently leaves Bob’s site open in her browser for the rest of the day (recording multiple user sessions and artificially boosting length of time on Bob’s site, both popular measures of online brand interaction).

Later Jane works her way through a pile of business-related direct mail offers — most of which get tossed in the trash unopened. On her commute home she passes the same billboards. Inside the front door she pauses to review the day’s mail (which includes a coupon for Bobs), before grabbing dinner (and a couple more of Bob’s commercials) and retiring.

Three week’s later Jane’s muffler gives out. She IMs her pal Sarah for a recommendation. Sarah tells Jane to check out Mandy’s Mufflers because it is female-owned and operated. Jane doesn’t even consider Bob’s before driving over to Mandy’s for an estimate.

Did Bob’s marketing work? Bob’s agency reached Jane nine times in just one day; perhaps two hundred times in the weeks between initial exposure and action.  The example is not too far fetched, as we see hundreds of ad messages every day.

Still Jane purchased elsewhere. So were any of the impressions decisive? It is doubtful that more ads would have changed Jane’s mind, but when exactly did Bob’s ads produce diminishing returns. What was going on in Jane’s mind as she cycled through awareness, consideration and purchase?  Why does she trust Sarah’s word more than Bob’s marketing? How could Bob spend his money more effectively? Should Bob put more online?

We have a flimsy sense of the complicated interplay of online advertising and overall brand awareness. Unable to match gross rating points and cost per thousand marketing metrics remain as far apart as oil and water.

Picture_21Above: The top video on YouTube on Saturday, September 30, 2006 had been viewed 92,357 times and had generated 155 comments and has been saved as a favorite 233 times (is this a measure of repeat customers?). With SuperBowl spots costing $ 3 mill each advertisers have every right to ask for similar levels of data from the mainstream media. 

Technology is charging into the chasm and will provide some answers soon. In stark contrast to Nielsen’s difficulty reporting the true number of people who watch TV commercials in the DVR age, YouTube and Google can both provide far more details about who watched what, for how long and when. New services like Slingbox can intercept traditional TV, rerouting the signal from a users cable box to their Laptop (gathering who knows what data for us to drool over). TIVO and NetFlix too are awash in consumer intelligence. And let’s not forget Web metrics companies like BlackFoot who seek to combine offline data from call centers and other commerce points  with online tracking data.

Few organizations can analyze total marketing consumption because it combines two different disciplines: the hard data of response metrics and the soft, qualitative feedback necessary to understand consumers’ emotional connection to brands and their less than rational patterns of behavior.

While we wait for technology to bring us more answers here are some of the simple, common sense guidelines the Mortar uses to answer questions of media efficacy and mix:

1. Start with tight goals.

What do you want to achieve? Drive sales? Improve internal buy-in? Strengthen loyalty? Introduce a new product? Boost market share? Spark discussion? The more specific your goal, the easier it will be to measure.

2. Define the customer.

Few marketers have a rich enough conception of their ideal customer. Consequently, at the beginning of every brand assignment, we push clients to seek alternative methods of defining their
customers. (For more read about our Mortar360 brand development process).

Picture_23Above: BlackFoot says your customers may be full of contradictions, but your data shouldn’t be.  Click through to see their new site (created by the Mortar).

If you are approaching your market with the same data as your rivals you will almost certainly reach similar conclusions.

Take Adult Education for instance. A quick Google search will provide a reasonably accurate demographic profile of potential students. And educational marketers often supplement their understanding with broad insights into motive such as career ambition or personal fulfillment. Yet, little of this is fresh or worse, unique to their student or institution. No wonder we are assailed by a large number of ads picturing grinning, smartly-dressed people leaping for joy because their degree has propelled them from the mail room to the board room.

Look for a new way to frame your audience. Great communications will follow.


Above: Click through to view this hilarious new "Got Milk" commercial from Goodby .

Goodby’s "Got Milk" developed from the realization that Mum’s buy most milk.

And that being without milk was far worse than buying too much. That insight provided a great campaign.

Online survey tools like Zoomerang and Survey Monkey can be quickly deployed to survey Mum’s about how much milk they intend to buy and if the "Got Milk" campaign had any impact — online or off.

Or Nike. "Just Do It" appeals to the athlete in all of us. Nike’s conception of their customer drove them to pitch their brand  as an enabler. "Just Do It" provides a platform that enabled them to move into the mainstream brand without impairing  credibility and attractiveness.

Returning to our example, let’s say Bob is anxious to counter the impact of Mandy’s Mufflers on his sales. Bob might convene some focus groups with women who have had muffler problems in the last 12 months and who have seen Bob’s ads but actually purchased at Mandy’s. A good planner can pull the answer out of in-depth explorations with surprisingly small numbers of people.

3. Ask "wassup?"

A tighter definition of customer limits the sample universe and allows marketers to develop a profile of common behavior.

For one of our clients we conduct annual surveys of working adults who have inquired and enrolled at the school. We look for changes in media consumption and trends that can be tracked back to promotion. Incidentally, our  surveys reveal more and more adult students are researching and applying for class online — prompting us to allocate more dollars to the Web.

4. Look to the long-term.

The Internet irreversibly collapsed marketing development schedules. What used to take months now takes days. Yes, we moan about it, but somehow we have made it work. And many of us are now secretly proud of our ability to execute effectively and at high speed. Nevertheless, marketing efficiency is best measured over the longer term. Look to develop plans across two or three year periods (gasp, yes, there is merit to planning longer than a Quarter Mr. Valley!).

5. And now for some heresy.

Although we risk heresy with the following statements experience prompts us to share these little nuggets:

  • With the absolute exception of pure direct response programs (like direct mail and online lead generation), if you are seeking an accurate measure of total marketing consumption numbers for a multichannel program including online and offline promotion and PR, you can rarely get to the numbers you need in less than three months. No matter how hard you try (unless you are Apple, in which case a couple of days is all you need). The Quarter is the absolute minimum required. So for all of you freshly minted MBAs hard at work on a pilot test for San Francisco, Chicago and Austin (you know who you are) you are best advised to scrap the offline components or extend the test period.
  • Very much the fashion in the last few years, Vanity URLs ( are losing their lustre and rarely produce resoundingly high amounts of traffic. And when they do work they produce results in a confusing, fragmented and resolutely roundabout manner. F’sure if Jane takes the time to fire up her browser and make her way to your custom site then she is certainly very interested in your products. Still its hard to measure the impact of what she does on mainstreet or what she says to her pals.   

Take "Subservient Chicken" which featured a man in a chicken suit willing to do just about anything users asked him too (along the way sparking no end of controversy over the effectiveness of viral marketing). That’s not to say Vanity URLs don’t belong in your aresenal of marketing tools. But leave room in your program for other results. For instance a visit to Subservient Chicken is only one way consumers show their appreciation for Burger King’s message. Offline word of mouth, increases in awareness or a shift in likelihood to purchase do not show up in Web traffic numbers. And unless your custom site features an online call to action (and you can’t buy burgers online), there is little (beyond an online coupon) to link offline behavior with Web reporting.

It is best to conceive of campaign-specific websites as one very visible part of the program–and steel yourself against criticism if the site doesn’t attract huge numbers of visits. The Web provides us with the quick hit of gratification. Reputations and campaigns often live and die by traffic numbers: even though we know better. Unfortunately, all too often the short-term indicators of success are misleading.

  • Conceive of your measurement program as a regular, annual activity and you can begin to really understand what your market is doing, and why. Put measurement in place at the beginning of your program and return to the market regularly with quantitative and qualitative tools to gauge how well you are moving the needle, and don’t be afraid to adjust spending accordingly.  Oh, and like us don’t be surprised if more and more of your ad dollars migrate to the Web.
  • And for those of you who still want that quick hit of instant traffic, fear not. Google and Yahoo have a keyword program for you right now. Grab your credit card and click through. If your site is ready to accept traffic you can have a marketing program set up in minutes. Kidding aside, keyword programs are amazingly effective. And they offer the most effective method of linking behavior with sales and lead generation. Unlike every other mechanism discussed here, you can have decent results from a keyword program in less than five days. Look to keywords to expand beyond lead generation in the next few years as marketers discover how to really leverage paid and organic keyword traffic.

Mortarmark welcomes your comments. You can reach him at here.

Post a Comment