Connecting the Data-Dots

Recently, my company sponsored me to trek into New York City for a day in order to emerge myself in the latest marketing technology trends at the ad:tech convention. The idea being that I could then report back with what I believe to be the most worthwhile findings. The event is, in fact, as advertised on their website: “an interactive advertising and technology conference and exhibition - a marketplace for buying and selling, a community for networking, a forum for exchanging ideas and an opportunity for contributing to industry trends and initiatives.”[1] ad:tech really is all of those things, but for my purposes, I made it a point to dive a little deeper into what the latest marketing attribution solutions available are. As I weaved my way from aisle to aisle, I took note of the vendors who haphazardly splashed the buzzword “attribution” on their booth, but could only give a superficial solution to a much more profound problem.

The problem I’m referring to is the daunting and often illusive task of connecting the data-dots. For marketers, the vast amount of data (from ever increasing Internet channels) that is provided online has made it difficult to determine which marketing efforts are really contributing to leads. And, as if connecting online data was not challenging enough, marketers need to consider connections between online and offline advertising as well. Consequently, although the data is abundant, many marketers still rely on old attribution methods (including last click) and are not adopting means of measuring the entire customer journey, via attribution models. Though most attribution models are considered imprecise, I believe that having one is still enormously beneficial. Businesses that use them find that attribution models undeniably lead to an increase in their ROI. I will use my trip to the Javits Center for the ad:tech convention to illustrate what I mean.

While in the city, one ad stood out to me in particular. It was an ad plastered on the side of a bus for the Radio City Christmas Spectacular. (I applaud Radio City for gaining my attention in what’s probably the most advertising saturated place a person can visit.) My manager and I noted that the models in this ad are intended to be seen as a family enjoying the show, but they didn’t particularly look related in our opinion. I realize that’s not a great reason for an ad to catch one’s attention, but still, it did. You can judge for yourself whether or not you think they look like a family as I have included the Facebook version of the ad below:

Figure 1: Radio City Music Hall Advertisement on Facebook.com

I kid you not, this ad appeared on my Facebook page the same day I saw the advertisement on the bus in New York City. I immediately began thinking about my online activity earlier that day. I looked up train times and I Googled the Javits Center. Could my apparent journey into the big apple be enough of a connection to link to a show? It’s either that or it was just a really great coincidence for Radio City to choose me to advertise to. Or they somehow peered into my mind and discovered that I am, in fact, a fan of the Rockettes (Obviously not the latter possibility.)

It’s either that or it was just a really great coincidence for Radio City to choose me to advertise to. Or they somehow peered into my mind and discovered that I am, in fact, a fan of the Rockettes.

Developing an attribution model may be beneficial, but the picture is never completely clear. That is to say, it is not an exact science, and what is effective for one campaign may not work as effectively for another. Wes Nicols, cofounder and CEO of Marketshare, wrote in his article Advertising Analytics 2.0: “That sort of insight represents the holy grail in marketing—knowing precisely how all the moving parts of a campaign collectively drive sales and what happens when you adjust them.”[2] Having some insight is better than having none at all. Much of that insight comes from taking the data we have and knowing how to use it. It’s no wonder, a bevy of marketing entrepreneurs have cropped up claiming to have the answers. What they mostly offer, however, are visually appealing dashboards that organize the data already available on Google Analytics.

Speaking of Google Analytics, they put together, in association with Econsultancy (a digital marketing agency in the UK) a report published in 2012 called “Marketing Attribution: Valuing the Customer Journey.” This paper is based on a survey as well as interviews conducted with marketers and agencies. Bill Key, the product manager of Google Analytics wrote of their findings, “Only 14% of respondents consider last click analysis (which, until recently, has been the industry standard) to be ‘very effective’. Yet over 50% of them are still using last click measurement—most likely because they haven’t yet found or mastered the right tools to take them beyond the last click.” [3]

This, in my opinion, represents a glaring shortcoming in the marketing industry. Marketers are supposed to be experts at leveraging data and using it to provide better results, and yet, there’s still a void in methodology. Relying on last click does not represent how marketing efforts work together and it does not give credit where credit is due. Those that attempt attribution models that combine data from multiple channels, however, see that it’s worthwhile.

Part of the reason marketers still use methods like last click is because the data part of the industry is rapidly evolving at a rate in which solutions and best practices have not yet surfaced. At a Salesforce conference I attended this year, they reported that 90% of the worlds data was created in the last two years.[5] Marketers are drowning is this data, but there are things that can be put into practice. Are these methods always accurate? No, not 100%, but these methods undoubtedly provide worthy results if they are used.

Are these methods always accurate? No, not 100%, but these methods undoubtedly provide worthy results if they are used.

The methods I’m speaking of come from marketer’s brains, not a software program. I’m talking about making judgment calls and analyzing all points of data in order to infer future marketing tactics, and through trial and error develop a marketing attribution model that is best for campaigns. One survey respondent said, “Attribution is complicated. Nothing gives you the full story, so you have to really understand your customers and assemble data and experience over time. We’re continually adding new layers of data and analysis methods, but it’s time consuming. It’s definitely worth it, but it’s time consuming and a work in progress for us.”[6]

If a little elbow grease and time are the main tools needed, that stills makes it hard to believe that 50% of marketers rely on last click. I would think that percentage would be less. The issue, as shown from the survey, has to do with company politics and money.[7] Companies are slow to adjust to new methods of marketing analysis, but they would if they didn’t feel there was so much risk. They feel that change could affect numbers negatively. As one survey respondent put it, “The elephant in the room is that attribution ultimately affects bonuses and the pockets of the people involved. You have to deal with the politics first or you won’t get buy-in, you’ll get resistance.”[8]

In many companies, marketing is segmented into traditional and digital workforces within their company, and unfortunately, those groups don’t typically work together. Since they don’t communicate, their data also does not communicate. This is a point of weakness when establishing attribution models. It would be interesting to know, for example, if the marketing department at Radio City Music Hall that allocates funds to plaster ads on buses correlates and attributes effectiveness of those ads with the ads fostered by their digital marketing department. (Note that I did try to contact their marketing team for an interview regarding these ads, but they did not respond.)

Nicols calls these marketing siloes “swim-lanes” and has suggestions on how to get out of those lanes. For one thing, he suggests that companies stop relying on the last-click method by drawing conclusions with the data they already have. In other words, stop viewing the performance of each marketing activity independently and start seeing how they correlate and assist one another. “Recognizing an assist depends on the ability to track how consumer behavior changes in response to advertising investments and sales activities. To oversimplify a bit: An analysis could pick up a spike in consumers’ click-throughs on an online banner ad after a new TV spot goes live—and link that effect to changes in purchase patterns. This would capture the spot’s “assist” to the banner ad and provide a truer picture of the TV ad’s ROI.”[9]

Microsoft is a great example of a company that was able to mingle their marketing channels when they began opening their brick and mortar stores. Initially, Microsoft didn’t sell directly to consumers, except for Office. A first version of Microsoft Store was online and eventually selling Office expanded into selling more products. Then five years ago, the brick-and-mortar stores were established.[10] This presented an interesting challenge for communication strategies and they realized their stores were competing with the online store, and information about the customers was not being shared.

Shawna Dahlin, Senior Email Marketing Manager at Microsoft, said, "As a big company, you would think that we have everything all in one place where we can act on it. But we unfortunately don't ... I'm sure there are lots of people that have that same issue," and added that when your data isn't integrated with your email platform, "all the fancy stuff that might get pitched to you from an email service provider, it's not easily done."

One of the main strategies that Microsoft used to deal with these issues was to develop and accept cross-channel integration. Although there was a data gap, the Microsoft team did have information about where customer had purchased their products, whether is be online or in a store. This key piece of information is how they leveraged their new strategy. Dhalin said: "As we're opening new stores, let's tell them about it … letting our customers decide where they want to shop. So, we added a dual call-to-action to every email. Let's have you find a store or shop online. You choose," she said. This was an effective leap from their previous method of having separate calls-to-action for the different purchase paths.

Microsoft, a huge corporation is the first to admit that it’s not an exact science, but it’s worth the time and effort to strategize. “It's not exact. We do our best with the data we have to tell them their closest store," Dhalin said. "It's helping drive that awareness across our channels of online and brick-and-mortar that you can shop anywhere, and we're going to help you regardless." The proof is in the pudding: Microsoft was able to increase their email revenue by 600%.

“It's not exact. We do our best with the data we have..."

An extremely vital task for marketers to do is connect the data-dots with reasonable inference, utilizing that information to make focused and logical strategies. As new data emerges, what must also emerge are new ways of using that data to better profit from what we know. Who knows, perhaps ten years from now I’ll be at ad:tech learning about new technology that will free marketers of having to use their own time and brain power. Until then, marketers must dive into the data and make the most of it.

Sources:

[1] ad-tech.com

[2] Nicols, Wes, (2013). Advertising Analytics 2.0, Harvard Business Review, 4

[3] Marketing Attribution: Valuing the Customer Journey (2012), Econsultancy.com, 1

[4] Marketing Attribution: Valuing the Customer Journey (2012), Econsultancy.com, 5

[5] Benioff, Marc, Salesforce World Tour, November 19, 2014

[6] Marketing Attribution: Valuing the Customer Journey (2012), Econsultancy.com, 9

[7] Marketing Attribution: Valuing the Customer Journey (2012), Econsultancy.com, 10

[8] Marketing Attribution: Valuing the Customer Journey (2012), Econsultancy.com, 10

[9] Nicols, Wes, (2013). Advertising Analytics 2.0, Harvard Business Review, 5

[10] Eckerle, Courtney. “Email Marketing: Microsoft Store uses relevance to increase sends by 300% and email revenue by 600%.” MarketingSherpa.com. Nov. 18, 2014

#article