Two questions to help you find the value of marketing


value of marketing

By Eric Wittlake, {grow} Contributing Columnist

I was just starting out in this marketing world and we were preparing to present the results of our first online advertising program to our largest client. We were thrilled. Our data clearly showed a big increase in the number of customers from advertising, and that our approach was significantly more cost effective than the prior agency’s. Life was good!

But during the presentation, we received a shock. Our client’s overall business had barely changed. Ouch!

It made me re-consider how we accounted for the value of marketing. We were counting, not causing. We knew people saw our ads and then became customers. But we didn’t have any idea how many of them would have become customers anyway. We had not measured our impact.

Re-thinking the value of marketing

Despite the advancements in data and analytics over the last 10 years, this same thing still happens every day. What’s worse, it is celebrated, with awards or case studies highlighting the success of things like social media programs that measure results through redemptions of a discount or special offer. Of course people redeemed … many were already regular customers!

Untangling causality in marketing is a complicated task that can take you down the road of marketing mix models, statisticians that speak a completely different language, and six-figure consulting fees. But for the rest of us, here are two simpler things we can do to start answering the question: In a given financial period, what was the value of marketing?

Track the right trend

Did the trajectory of your business change? If you can’t prove a relationship between marketing and business success, don’t expect a high-five from the CEO and a big bonus.

Revenue isn’t the only useful trend line. What’s important is to look at the metrics for the entire business, not only a marketing-attributed view. In the cyclical and long-sales-cycle world of enterprise B2B sales (where I spend much of my time today), these metrics are often more helpful than revenue trends:

  • Branded searches. How many people are searching for your brand or product by name? Impressions and clicks on your brand terms (available in Google Webmaster Tools) are a useful metric and one of my favorite gut-checks of impact.
  • Direct website visits. Focus on direct visits only (and make sure any paid media is being coded to easily be identified and tracked!)
  • Market share. If your marketing is really driving the change, you will be winning market share, not just growing with your market.

Of course, a trend line doesn’t mean marketing is driving the change. A product launch, a change in customer service, or the collapse of a competitor can confound the data. But you should be applying statistical analysis and charting business changes to tease out the truth as best you can. Given a long enough period of time, you normally can demonstrate the relationship between marketing and these trends.

Test the trend

How much of the change in your business is really because of marketing? In the short-term, there might be ways to test the effect.

You will need to divide your audience into two similar groups. Roll out your marketing to one group, exclude the second, and then see how your business does within each group. Is the group that sees your marketing buying more or buying more frequently? Or do the two groups seem to be about the same?

Depending on the kind of marketing you are doing, here are four ways you can consider creating these test and control groups.

  • Isolate individual markets. If you are launching a new advertising campaign, introducing new product packaging or testing a promotional offer, start by rolling it out in just a few cities or DMAs and compare the results to those from the rest of your market. Even better, compare the results to those in similar markets (sometimes referred to as matched market testing).
  • Split a company list. If you market to specific account lists, either through email and direct mail, or through one of the newer company-targeted advertising offering, leave a few target companies out of your marketing and compare win rates or even traffic to your site from each group.
  • Create a digital advertising control group. If you are primarily running or testing online advertising, carve out a random sample of people and exclude them from your campaign. This is the approach we took 10 years ago and we learned that less than 5% of the results we reported (or counted) were actually caused by marketing. I expect to see more of this in the future, driven in part by Facebook’s move to allow advertisers to track campaigns using Facebook’s identifier instead of a tracking cookie.
  • Vary across time. Although it isn’t ideal, if content marketing or social media are major components of your marketing mix, either turning them on-or-off, or simply varying the level of activity month-by-month, can yield valuable insights. If you take this approach, plan to repeat the on-and-off cycle a couple of times before drawing any conclusions.

Brace yourself. In my experience, as you dig into marketing performance, you will nearly always find that the real impact of marketing is significantly less than what you have been sharing in nicely polished performance reports. But I’d rather know the real result than live in a my own little marketing fantasy land. What about you?

Photo Credit: Ervin Bartis via Flickr cc

Eric WittlakeEric Wittlake spends his days working with B2B marketers and (occasionally) shares his marketing views on his personal blog, B2B Digital Marketing. You can find him on Twitter (@wittlake) when he isn’t working with B2B marketers.

All posts

The Marketing Companion Podcast

Why not tune into the world’s most entertaining marketing podcast!

View details

Let's plot a strategy together

Want to solve big marketing problems for a little bit of money? Sign up for an hour of Mark’s time and put your business on the fast-track.

View details