Marketing Measurement and the Mississippi River
By Tim Wilson on in Facebook Insights, Metrics, Web Analytics with 2 Comments
At least once a week in my role at Resource Interactive, I get asked some flavor of this basic question: “How do I measure the impact of my digital/social media investment?” It’s a fair question, but the answer (or, in some cases, the impetus for the question) is complicated and, often, is related to the frustration gap — the logical leap that, since digital marketing is the most measurable marketing medium of all time, it enables a near-perfect linkage between marketing investments and financial results.
It’s no fun to be the bearer of Reality Tidings when asked the question, especially when it’s easy to sound like the reason we can’t make a clean linkage is because it’s really hard or we just aren’t smart enough to do so. There are countless sharp, well-funded people in the marketing industry trying to answer this exact question, and, to date, there is a pretty strong consensus when you get a group of these people together:
- We all wish we had “the answer”
- The evolution of consumers and the growth of social media adoption has made “the answer” more elusive rather than less
- “The answer” is not something that is just around the corner — we’re chipping away at the challenge, but the increasing fragmentation of consumer experiences, and the explosion of channels available for marketers to engage with those consumers, is constantly increasing the complexity of “the question”
That’s not an easy message to convey.
So, How’s That Explanation Working Out for Ya’?
It’s a tough row to hoe — not just being a data guy who expends a disproportionate amount of energy, time, and brainpower trying to find a clean way to come at this measurement, but trying to concisely explain the complexity. Of late, I’ve landed on an analogy that seems to hold up pretty well: measuring marketing is like measuring the Mississippi River.
If you are tasked with measuring the Mississippi, you can head to New Orleans, don hip waders, load up a rucksack with instruments, and measure all sorts of things at the river’s mouth: flow volume, fish count, contaminants, etc. That’s analogous to measuring a brand’s overall marketing results: brand awareness, share of voice in the industry, customer satisfaction, revenue, profitability, etc. The explosion of digital and social media actually makes some of this measurement easier and cheaper than ever before through the emergency of various online listening and social media analytics platforms.
While these “mouth of the river” measures are useful information — they are measures of the final outcome that really matters (both in the case of the Mississippi and brand marketing) — how actionable are they, really? As soon as results are reported, the obvious questions come: “But, what’s causing those results?”
What causes the Mississippi River to flow at a certain rate, with a certain number of a fish, with a certain level of a certain contaminant where it empties into the Gulf of Mexico? It’s the combination of all that is happening upstream…and the Mississippi’s headwaters reach from Montana (and even western Canada) all the way to Pennsylvania! The myriad headwaters come together many times over — they interact with each other just as different marketing channels interact with and amplify each other — in thousands of ways over time.
If we’re looking to make the Mississippi cleaner, we could travel to western Kansas and check the cleanliness of the Smoky Hill River. If it’s dirtier than we think it should be, we can work to clean it up. But, will that actually make the Mississippi noticeably cleaner? Logic tells us that it certainly can’t hurt! But, rational thought also tells us that that is just one small piece in an almost incomprehensibly puzzle.
With marketing, we have a comparably complex ecosystem at work. We can measure the growth of our Facebook page’s fans, but how is that interacting with our Twitter feed and our web site and our TV advertising and blog posts that reference us and reviews of our products on retailer sites and our banner ads and our SEO efforts and our affiliate programs and our competitors’ presence in all of these areas and… ugh! At a high level, a marketer’s Mississippi River looks like this:
Not only does each of the “managed tactics” represent dozens or even hundreds of individual activities, but environmental factors can be a Mack truck that dwarfs all of the careful planning and investment:
- Cultural trends — do you really think that the Silly Bandz explosion was carefully orchestrated and planned by Silly Bandz marketers (the CEO of Silly Bandz certainly thinks so — I’m skeptical that there wasn’t a healthy dose of luck involved)
- Economic factors — during a global recession, most businesses suffer, and successful marketing is often marketing that manages to simply help keep the company afloat
- Competition — if you are a major oil producer, and one of the top players in your market inadvertently starts dumping an unfathomable amount of crude into the Gulf of Mexico, your brand begins to look better by comparison (although your industry as a whole suffers on the public perception front)
“It’s complicated” is something of an understatement when trying to accurately measure either the Mississippi River or marketing!
So, We Just Throw Up Our Hands and Give Up?
Just because we cannot practically achieve the Holy Grail of measurement doesn’t mean that we can’t be data driven or that we can’t quantify the impact of our investments — it just means that we have to take a structured, disciplined approach to the effort and accept (and embrace) that marketing measurement is both art and science. In the Mississippi River example, there are really three fundamentally different measurement approaches:
- Measure the river where it flows into the Gulf of Mexico
- Measure all (or many) of the tributaries that feed into each other and, ultimately, into the main river
- Model the whole river system by gathering and crunching a lot of data
The first two approaches are reasonably straightforward. The third gets complex, expensive, and time-consuming.
For marketers — and I’m just going to focus on digital marketing here, as that’s complex enough! — we’ve got an analogous set of options (as it should be…or I wouldn’t be calling this an analogy!):
Measuring the direct and cross-channel effect of each tactic on the overall brand outcomes is nirvana — that’s what we’d like to be able to do in some reasonably reliable and straightforward way. And, we’d like that to be able to factor in offline tactics and even environmental factors. For now, the most promising approach is to use panel-based measurement for this — take a sufficiently large panel of volunteers (we’re talking 10s or 100s of thousands of people here) who voluntarily have their exposure to different media tracked, and then map that exposure to brand results: unaided recall of the brand, purchase intent, and even actual purchases. But, even to do this in an incomplete and crude fashion is currently an expensive proposition. That doesn’t mean it’s not an investment worth making — it just means it’s not practical in many, many situations.
However, we can combine the other two approaches — measurement of tactics (tactics include both always-on channels such as a Facebook page or a web site, as well as campaigns that may or may not cut across multiple channels) and measurement of brand results. The key here is to have clearly defined objectives at the brand level and to align your tactic-level measurement with those same objectives. I’m not going to spend time here expanding on clear definition of objectives, but if you’re looking for some interesting thinking there, take a look at John Lovett and Jeremiah Owyang’s white paper on social marketing analytics. They list four basic objectives that social media can support. At the overall brand level, I think there are basically eight possible objectives that a consumer brand might be tackling (with room for any brand to have one or two niche objectives that aren’t included in that list) — and, realistically, focusing in on about half that many is smart business. But I said I wasn’t going to expand on objectives…
What is important is to apply the same objectives at the brand and the tactic level — each tactic isn’t necessarily intended to drive all of the brand’s objectives, so being clear as to which objectives are not expected to be supported by a given tactic can help set appropriate expectations.
Just because the objectives should align between the tactic and the brand-level measurement does NOT mean that the measures used to track progress against each objective should be the same. For instance, if one of your objectives is to increase engagement with consumers, at the brand level, this may be measured by the volume and sentiment of conversations occurring online about the brand (online listening platforms enable this measurement in near real-time). For the brand’s Facebook page (a tactic), which shares the objective, the measure may, instead, be the number of comments and likes for content posted on the page.
But…How Does That Really Help?
By using objectives to align the measurement of tactics and the measurement of the brand, you wind up with a powerful performance measurement tool:
As simplistic and extreme examples, consider the situation where all of your tactics are performing swimmingly, but the brand overall is suffering. This might be the result of a Mack truck environmental factor — which, hopefully, you are well aware of because you are a savvy marketer and are paying attention to the environment in which you are operating. If not, then you should consider revisiting your overall strategy — do you have the wrong tactics in place to support the brand outcomes you hope to achieve?
On the other hand, consider a situation where the brand overall is suffering and the tactics as a whole are suffering. In that case, you might have a perfectly fine strategy, but your tactical execution is weak. The first order of business is to get the tactics clicking along as designed and see if the brand results improve (in a sense, this is a preferable situation, as it is generally easier to adjust and improve tactics than it is to overhaul a strategy).
In practice, we’re seldom working in a world where things are as black and white (or as green and red) as this conceptual scenario. But, it can certainly be the case that macro-level measurement of an objective — say, increasing brand awareness — is suffering while the individual tactics are performing fine. Let’s say you heavily invested in your Facebook page as the primary tactic to drive brand awareness. The page has been growing total fans and unique page views at a rapid clip, but your overall brand awareness is not changing. You may realize that you’re starting from a very small number of fans on Facebook, and your expectation that that tactic will heavily drive overall brand awareness is not realistic — you need to introduce additional tactics to really move the brand-level awareness needle.
In the End, It’s Art AND Science
Among marketing measurement practitioners, the phrase “it’s art and science” is oft-invoked. It sounds downright cliché…yet it is true and it’s something that many marketers struggle to come to terms with. Look at marketing strategy development and execution this way:
“The data” is never going to generate a strategy — knowing your customers, your company, your competition, and a bevy of other qualitative factors should all be included in the development or refinement of your strategy. Certainly, data can inform and influence the strategy, but it cannot generate a strategy on its own. Performance measurement, though, is all about science — at its best, it is the quantitative and objective measurement of progress towards a set of objectives through the tracking of pre-defined direct and proxy measures. Dashboards can identify trouble spots and can trigger alerts, but their root causes and remediation may or may not be determined from the data — qualitative knowledge and hypothesizing (“arts”) are often just as valuable as drilling deeper into the data.
It’s a fun world we live in — lots of data that can be very valuable and can drive both the efficiency and effectiveness of marketing investments. It just can’t quite deliver nirvana in an inexpensive, easy-to-use, web-based, real-time dashboard! 🙂