Like many marketers, I have fallen in love with the characters, the pacing and the work environment of the AMC Original Series, Mad Men. One of the things that amazes me though, is how little attention is placed by clients or Sterling Cooper Draper Price employees on getting a return on investment (ROI). Sure Rachel Menken talks about her need to increase revenue at the department store she runs; and Peggy creates the “Our hams are worth fighting for” guerilla marketing campaign for Sugarberry Ham. What’s lacking, however, are the numbers. In my world, my typical day at work, projections and actuals of customers, subscribers, cancellations, and leads are central to the conversation.

No matter how good my ideas are, before I can hire a couple of actresses to fight over a ham, let’s say at Georgetown’s high profile “Social Safeway”, I need to build a business case and I need to stake my efforts on projections that speak to the bottom line. In many ways, this is frustration. It limits creativity and slows down the well-meaning efforts and activities of marketers and even agencies. But we do this because putting market research and financial modeling in the front seat, is supposed to help us make more efficient use of our time.

In case you haven’t noticed, there’s a big problem with this calculation. If we were willing to settle for the standards of success that prevailed in Mad Men’s 1960s, much less labor would be required. If your bosses are anything like mine, we’re not! Instead, we layer on requirements to get ideas and projects through. In our noble efforts to be MORE efficient with our time, we end up creating much more expense and cost on the front end of our campaigns and we created a world where marketing – as Betty Friedan so famously said about housewifery in “The Feminine Mystique” – expands to fill the time available.

Clearly one of the biggest culprits in the expanding workload challenge is social media and more generally word-of-mouth marketing. And yet neither on or offline word of mouth can be ignored. Here’s what McKinsey genius Jacques Bughin and his co-authors claim in their April 2010 piece in the McKinsey Quarterly:

“Word of mouth is the primary factor behind 20 to 50 percent of all purchasing decisions. Its influence is greatest when consumers are buying a product for the first time or when products are relatively expensive, factors that tend to make people conduct more research, seek more opinions, and deliberate longer than they otherwise would.”

Word of Mouth is important, but its’ effect is about as difficult to measure as Peggy’s efforts to start a ham fighting trend. The complexity of communities and of individual behavior have lead to some standard short cuts. More than once, I’ve been pitched software that would scour the Internet for mentions of my company or product and count those mentions as a measure of social media success. Some of these tools even attempt to classify mentions as positive, negative, or neutral though I’ve never been impressed with the algorithm’s classifying ability.

Bughin admits: “There’s an appealing power and simplicity to this approach” but, he points out, when we are counting mentions, even if we are categorizing them accurately as positive or negative, we are leaving out a key – perhaps THE key – differentiator between the messages.

Research shows that it’s not just what is being recommended or the number of times something is being recommended that influences adoption rates. The person behind the recommendation plays a significant role. A recommendation made by a family member has far more influence than one made by a stranger.

As Bughin writes, “Our research shows that a high-impact recommendation—from a trusted friend conveying a relevant message, for example—is up to 50 times more likely to trigger a purchase than is a low-impact recommendation.”

So McKinsey has given market researchers some new guidelines for evaluating word of mouth messages. I haven’t put this to work yet but I have a campaign coming I am going to try this with. I’d love to hear your thoughts.

Here’s my plan:

Step 1

Create a content sample of at least 100 messages. I have a researcher who will create a list ordered by date of all the mentions we know of over a 3 month period. I’ll have her enter them into an excel spread sheet. Let’s say there have been 1,000 mentions in Q3 that we can find. In that case, to get a “random” sample of 100 messages, I’ll just select every 10^th message and delete the rest. It’s not going to be a statistically significant sample, but I do want it to be random so all the messages aren’t on a certain topic or from a certain date.

Step 2

Categorize all 100 messages into buckets as laid out by the McKinsey report. One message, by the way can be in multiple buckets. Here’s how Bughin says to organize them:

Bucket A: Messages that address important product or service features. [This one surprises me since I do make a concerted effort to position products emotionally, but McKinsey says “consumers actually tend to talk—and generate buzz—about functional messages.”]

Bucket B: Messages from people consumers would be likely to trust and believe that he or she really knows the product or service in question. [McKinsey says about 8 to 10 percent of consumers are what we call influentials, whose common factor is trust and competence. Since I can’t identify friends and family of readers, I am sticking with celebrities, business owners, and experts in the category as a proxy.]

Bucket C: Messages passed within tight, trusted networks. [This isn’t going to be easy but when we put together our list of Q3 mentions I am going to review the customer service records and email correspondence for evidence of face-to-face recommendations in small groups and between family members. For instance the product I am going to use this for has an exclusive “Friends and Family” discount and I am going to treat each of those sent emails as a word of mouth message.]

Step 3

Once I have the sample categorized I am going to apply McKinsey’s measure of word-of-mouth equity model below to create an influence chart which “represents the average sales impact of a brand message multiplied by the number of word-of-mouth messages.” And I’ll look at sales numbers to apply a dollar value to this impact.

Here’s what McKinsey says my chart should look like:

This is a much more intelligent way of looking at word of mouth and it addresses some of the short comings are facing. Of course the biggest issue for many small companies – and I suspect plenty of the big ones too – is being able to capture and categorize the messages, particularly off-line ones. McKinsey doesn’t offer much hope or help there and ultimately that’s where a lot of time and expense can occur. It’s also where you can inadvertently alter your results because of inaccurate sampling. But in principal, Bughin’s heading in the right direction and he’s got my attention.

As marketers, we are well aware word of mouth is probably the most significant tool in our arsenal. That’s not new, as witnessed by the Sugarberry Ham example set in the 1960s. But for too long we have used the difficulty in measuring results for not bothering to try. Those heady days are long over. Upper management, boards, and shareholders will no longer stand for it. So while word of mouth requires creativity for success, we aren’t going to get the funding to be creative without putting some more science behind the art. Marketers need to step out of their comfort zones and start estimating the ROI of word of mouth.

Read the McKinsey article, A new way to measure word-of-mouth marketing, in full here: https://www.mckinseyquarterly.com/article_print.aspx?L2=16&L3=20&ar=2567