Analytics is complicated and, because it is designed to provide facts, is not often as welcome in a marketing discussion as other disciplines. Analytics sits between marketing and technology in a place where neither party feels altogether comfortable; and too often the analytics consultant finds herself taking arrows from both sides.
Here are some suggestions for the analyst in order to make sure your analytics customers remain happy, whether internal or external.
1. Establish Legitimacy (Confidence)
Do you know what you are talking about? Make sure your customer knows it. Often the more insecure a stakeholder is, the more skeptical they are of you. When you meet these kind of folks for the first time, have handy a brief “elevator pitch” that lists your qualifications and experience in a friendly way. Then, avoid jargon and instead, talk in a way that some folks call “storytelling” but which I call “narrative.” This requires you know the data and where it points. In this scenario, you are the scout-leader heading the group on a hike to the facts.
Don’t make yourself a burden by going over the top, but when you wonder “should I check in,” it often means your subconscious has already answered that question and is trying to get your attention. Making sure your stakeholders are nearly as well-informed is key to keeping them happy. Communicate more judiciously than you would with a colleague. Your customer is more averse to surprises, and that’s mainly because they often have their own reports to do, and when you surprise them, then they have to surprise their boss. And their boss really does not like surprises. Keep from surprising your stakeholder, even when you think the news is good.
3. Test Before Launch
Have you heard of the “small technical glitch” that “caused a big problem”? It happens a lot more than you’d expect. Almost always, this is because no one has set up a proper testing environment; and tested whether the program creates unexpected changes in data collection or reporting. A test environment is a great way of avoiding surprises (see above). In many cases, it can spell the difference between a good analytics program and loss of confidence.
4. Pay Attention to Narrative
People are storytellers. Data doesn’t tell a story, but it does provide you with reports so that you can tell a story. Perhaps one day there will be a truly engaging storytelling robot but today, it is still the job of the human being to look at seemingly unconnected threads of information, see patterns, understand nuance and relative importance, and to create a story out of raw numbers. You’ll likely need data from different sources to create a fully dimensional picture for yourself, which then you will use to create the narrative that comprises the insight needed to make changes based on data. Without the narrative, it’s just machines talking to other machines.
5. Don’t Defend Technology at the Expense of Business Needs
Technology is not business, it is a subset and a provider to business. So when a non-technical person says why not, the technologist is ill-advised to simply say “we can’t do that,” assuming the non-techie will accept that “the technology just cannot do that.” First, you may not be right. Very often, there is a solution out there, and maybe you need to find it. Second, many businesspeople see a technology lack as your lack, because without technology, you would not be there at all. It’s OK to say the technology cannot do it if the technology cannot do it, but you will need to communicate that as a business concept. For instance, “there is no data source” is not nearly as effective as “we need someone to give us access to the data sources, do you know who that might be?” All of your reasons for doing things (or not doing things) must serve a business purpose, or you need to supply a plausible business reason why you can’t do it.
With these five weird tricks you should be able to lose weight, get cheap car insurance, and even keep your analytics customers happy!
According to recent Nielsen numbers on the topic of Internet ad-spend versus TV ad-spend, the Internet is like a small bird picking mites off the back of an elephant. I will admit these numbers are nine months old, but I do believe they still pertain, even as mobile ad-spend has pushed Internet spend higher in the last couple of years.
Nielsen gives less than 6 percent to “on line” while “television” (which includes cable) gets close to 60 percent of total global ad spend (including all advertising everywhere). If you think broadcast is still relevant, it might interest you to know that the Internet has already surpassed it. And while TV has grown much slower than the Internet, the Internet is still very small.
While Nielsen does not publish dollar amounts, ZenithOptimedia says the global total ad-spend surpassed half a trillion dollars in 2013. So both the little bird and the big elephant are billion-dollar babies. Zenith also gives the Internet a whopping 21 percent of spend, still dwarfed by TV. This may be because Nielsen gives no credit to search-engine ads, only display ads. Which is only a little bit silly, because last I read, Google has more ad-spend than all of print media in the U.S.
The elephant and the little bird are both “in the room,” if you will.
What’s the Hold-Up?
Why is TV still dominant? I read somewhere we had stopped watching TV; so I must have missed something.
In light of all the above, why is Internet the darling and TV too often portrayed as somehow the inevitable loser in this race?
If you want to hear answers from the tech community, it’s a mere statistical anomaly that the Internet has not yet completed its domination; and that advertisers just need to get with it. If you want to hear answers from advertisers, especially those that control the massive TV ad budgets, they may tell you it’s because when they go to buy ad space on the Internet, they don’t know what in the heck they are buying.
TV ad spend is pretty much ruled by Nielsen and that is because it always has been that way, and it is an article of faith that Nielsen knows best.
However silly that might be, the fact there is a single source of well-defined audience measurement for TV makes all the difference.
Digital analytics, which is supposed to be much more accurate than the extrapolated data from Nielsen, cannot begin to compete; not least because it is like a mirror shattered into a million pieces — each shard with its own reflection — while Nielsen is just a single mirror where you can see stuff.
So, is it really about slavish fealty to Nielsen? Or is there some other reason why Internet advertising continues to lag despite robust year-over-year growth?
Maybe It’s the Content
Is it possible the answer is that no one can stand Internet ads? And that they often refuse to click on them? I have one friend who says that whatever brand shows up in his free mail account is the brand he will not buy. He never said that about the ads that interrupted the ballgame.
I think there are natural limits to every market, and that Internet advertising may soon find that limit. It isn’t because people don’t love the Internet. It’s because when they watch TV, some commercials actually get their attention and make them laugh or listen. I have some doubt as to whether anyone anywhere has ever said to themselves, “Now that’s a cool Internet ad.”
I don’t have the answer — and I don’t click on ads either. But I do like that ad on TV with the hamsters driving a little Korean car.
You shouldn’t be.
In fact, if your digital footprint is large and complex, you ought to be clamoring for it.
That such an arcane digital offering should in fact be a bedrock of data collection success is only testament to its primacy.
Why should you care?
Let me first state the problem:
In a galaxy far, far away, perhaps they continue to use server log files to figure out what users did on a website, but in that same galaxy they are probably still experimenting with more effective ways to build a fire with two sticks and no matches. On Earth, almost no one uses log files anymore.
We deploy digital analytics tools that rely on a much more targeted and accurate manner of collecting user data: tagging.
Why is this such a big deal, and why is improper or nonexistent tagging the most common stumbling block to reliable analytics?
It’s a big deal because tags enable (relatively) precise data collection. It’s also a big deal because getting tags properly implemented and then managed in an enterprise is about as easy as herding jungle cats – the kind that maul you suddenly.
Without tag management, “the process” (more accurately “the chaos”) almost invariably looks something like this:
A marketer says they need to know X, Y, and Z about user behavior on a digital property. Typically it goes beyond the basics of unique visitors and total page views. In order to get reports that illuminate user behavior, the marketer needs someone to create a “tag specification”: a document, often a spreadsheet, that describes the reporting need and the tag (small snippets of code) that must be placed in HTML so that when reports are created, they in fact have data in them. This part is usually executed without too much difficulty as long as the tag-specifier knows the tools and the tags very well.
Then, too often in my experience, the chaos really begins.
That’s because people who create tag specifications do not control (and do not want to control) the HTML that drives the site. The tagging spec needs to be handed off to developers (not the marketers) so they can place these tags. Sometimes it goes OK. Often it does not. Tags are placed, but incorrectly. Or they are not placed, and with little explanation as to why they are not. Weeks go by. Marketers get frustrated and too often end up settling for much more basic reporting than they had hoped. Often they end up with whatever basic reporting comes from putting just the simple tag provided by the vendor and the tagging spec becomes an artifact of hope but with little chance of becoming a robust reporting suite.
Compounding the problem in an enterprise is the multitude of sites that have wildcatted their own analytics, usually flawed implementations of free tools like Google Analytics (which can be quite powerful when properly deployed). Nothing can be measured against anything else because there is no common tagging protocol, no agreement on what is measured and how it is reported upon.
Tag management solves both of these problems rather handily and should be part of every marketer’s toolkit where they have more than a couple of sites to manage.
There are several flavors of tag management tools, ranging from Tealium’s container paradigm to Ensighten’s “server conversation” to Adobe’s integrated package and more. What they all have in common is inherent in the category name: They manage tags and tagging.
With tag management, instead of sending off a tagging spec to a range of developers responsible for different sites hoping they all do the tagging in the same way, the marketer can have one group of developers implement a tagging structure once, and then deploy it widely.
Moreover, they can, using a graphical user interface (anyone remember the term “GUI”?), decide the conditions under which certain tags “fire” (become active); and on which digital assets. Parameters can be set globally and then with little fanfare, the tag management system actually manages these tags globally. With robust error-detection and selective deployment capabilities, many if not most of the problems associated with data collection simply go away.
Tag management tools centralize the management of data collection and can be the foundation of excellence in reporting. It’s really that simple.
You ought not be bored by tag management. You ought to be excited about it, you ought to deploy it, you have every right to benefit from it once deployed. There is little excuse today for an enterprise not to use tag management in maintaining control over data collection and reporting over a wide range of sites.
If you’re still wondering why your enterprise analytics are in disarray, you now know why. Don’t let a good tag specification go down to defeat. Collect data reliably and globally.
With tag management.