Analytics at the Marketing/Finance Interface

Jim Novo is right about the relevance of financial data to marketers. There's a wealth of information in those transactional records, and most companies don't have analytics practitioners embedded within those departments.

Just as reporting (reportage) isn't analytics, accounting isn't analytics. There's something to be said about the parallels between finance and analytics. Both have prediction, definitions, jargon, simulation, machine learning, ratios, rules, laws, and statistics. Both become confused with bookkeeping and dashboarding. One is older than the other.

If accounting/finance is an 110 year old, marketing analytics is 18.

There are very specific rules in bookkeeping, handed down for centuries. There are very specific standards in finance, handed down since at least the Renaissance. Should we be so lucky in marketing reporting, web analytics, analytics, and marketing science.

(Indeed, the very fact that marketing science has to add the word 'science' says something about how young the field is (founded in 1961). It's called 'chemistry', not 'chemical science', after all.)

So there's that axis and distinction - between counting and reporting, and analysis and prospection. Pretty solid ground there.

Marketing analytics becomes more complex because of the mixing of transactional data with attitudinal/behavioral data.

There's a whole set of data that most finance people don't have to think about. Certainly, in certain spheres of finance marketing, 'psychology of the market' starts to play a role. Those are some pretty early days. I've met only two of them. And it would appear that they're in a super minority.

Jim Novo, in a previous comment thread, wrote: "Marketing history is full of examples where Awareness did not turn into revenue."

I can think of many investments that did not turn into revenue as well.

It would appear that both disciplines are united in many ways.

The professional who tries to bridge this chasm, between marketing and finance, can be successful. But they'll have to recognize two things:

• They will have to make the distinction between accounting/reporting and finance/analytics very quickly.

• They will have to make the distinction between transactional data, behavioral data, and attitudinal data, very quickly.

It's worth pursuit.

How much are hackers saying about data and analytics?

Some excellent work by Jerod Santo and his Hacker News Trend visualizer.

Pretty sweet.

Hacker News is a nice, social curation program. It's a pretty good indicator of what data developers and hackers are thinking and doing in many markets. It's one of the main bellwethers of the bootstrap community, and there's typically a lot of content that I find interesting.

All trend lines that follow are adjusted for size of community at the time.

The first chart, below, describes the persistence of analytics and the rise of specific, applied uses - the a/b test. It also shows that the term 'data science', as a category of interest, had a bit of a bubble in late 2009.

That's great, but how about some perspective? What are some of the biggest topics that dominate HN?

Check the Y-axis. We're talking about much higher volumes. Note the rise of apple and the decrease in hack. Mainstays 'work' and 'show', two keywords I pulled from some of the excellent work MetaMarkets did, hold steady.

Note the rise of Apple and the rise of Data.

I've sandbagged a cluster - outputs of data versus the term data itself.

What do I see?

I see the rise of data, and then flatness. A natural ceiling at sub 10,000 where, possibly, the community won't tolerate too much of the same thing without becoming OJ'd/Weinered/Lybianed out. Ie. Topic burnout (see the previous charts). It may also be caused by something more sinister.

The rise of data always leads to the dissatisfaction, because data doesn't answer, all by itself:

• What does it mean,
• What does it mean to me,
• What does it mean for my future, and
• How can I use it to make my future better.

It may be dissatisfaction.

Analytics professionals know how this goes. If there was ever a time to connect marketing scientists, web analysts, and data scientists with hackers and developers, it should be right about now.

Ideally, we would see more applied analytics terms rise from below, while data begins its slow, inevitable, decline.

In the absence of statistical analysis - am I being fooled by randomness here - or is there really something here?

Analytics at the Marketing/Technology Interface

Analytics at the marketing / technology interface. What a jam.

On the one side, there are massive time constraints on developers who struggle with super tight deadlines and last minute tweakings. On the other side, you have marketers who are buffeted by super tight deadlines, little time to plan, and tend to follow an anchor-and-adjust mentality. (Don't we all?).

Much of the frustration within analytics derives from watching both sides of the equation, and actively attempting to collaborate between two groups. I can sum up the fight between the two as:

Marketer: "You go too slow!"
IT: "You tell us about a project at the last minute!"
Marketer: "You're too slow!!!"
IT: "You don't know how to make choices!"
Marketer: "You're too slow!!!"
Analytics: "Guys, why are my tags broken?"

A few thinkers believe that analytics would find a better home within Finance. And perhaps certain aspects of marketing analytics would find a better home in Finance. I don't really think so at this point. Consider the following:

Finance is transactional. They have clean data, and for the most part, except for creative accounting, they have black and white distinctions. They've been able to milk very simple formulas for a very long time.

Marketing is both transactional and attitudinal, and this mirrors much of the Marketing/Technology interface.

When we're dealing with technologists, we're dealing with the transactional. And there are very specific ways that data is ordered, counted, and segmented. Those who are more technical tend to have more in common with finance. Look at how much you can milk the visit metric. It's beautiful.

When we're dealing with most marketers, we're dealing with the attitudinal, and in many circumstances, the transactional.

How many people really loved that microsite I spent forever on?
How many people really loved that post?
What percentage of my target is aware of that commercial I put out?
Did it affect the way people think of us?
Did it affect the way people make decisions?
Did it change purchasing behavior?

These are much richer and problematic questions.

It's the linkage of what a marketer is doing to a system (cause) to the effect of what can be observed in that system. It would be easy except that that system is dynamic and chaotic. And it's not all purely transactional.

So, yes, to my friends who are technologically oriented, yes, I think we have a long ways to go in solving for the technological issues underlying transactional measurement. And those challenges are by no means easy. (I'm living through that right now. It's not easy!). And perhaps the linkage between finance and technology, in assessing financial marketing performance, is the sum of all fears for certain marketers. Maybe that fusion needs to happen.

I don't advocate for all of us going to finance because it leaves the attitudinal on the table.

I think we have an even longer way to go in linking the attitudinal data with transactional data. And those challenges are just as technical as they are human. That is to say, we're going to have to invent a new schemas for thinking about choice, preference, distribution, price, and social that agree with actual observation. This is where marketing science has gotten it right and economics has gotten it wrong. Economics has become, largely, math and assumptions for the purpose of math. Marketing science has become models and math for the purpose of prediction. Marketing scientists are trying it. Analysts, in pockets around "VOC" or "website satisfaction", are too.

This is where a certain degree of analytical professionalism is going to have to come in. If analysts really want to rise to this challenge and try it, they're going to have to accept that two models may both indeed be valid. They're going to have to accept that two people can be simultaneously equally right, or equally wrong. Or one may have a simpler explanation that is more memorable than another more complex one, yet one is more predictive.

Life at the marketing/technology interface is a lot like a large mass of water trying to flow through a very small aperture. There's a whole lot of spin and there's a whole lot of energy expended to just get it jammed through.

There's a gathering in Toronto this Wednesday. The theme of the night, "Does Anybody Give A Shit About Your Dumb Idea?", and it is directly at the marketing/technology interface - combining the attitudinal with the transactional. I hope to see you there.

Are you Analytical Ketchup, Mustard, or Kobe Beef?

Ketchup, in North America, generally comes in one flavor. It's that familiar sugary taste with a hint of what is supposedly tomato. It doesn't really taste like a tomato though. It's different elsewhere, like in Germany, where it seems to be more juicy. But in North America, there's one generally accepted flavor.

People know what they're getting with ketchup. They know how much a bottle of it will cost at the costco, the grocery store, or the convenience store. They have a firm idea of what to expect. It's ketchup.

There's a general expectation around ketchup. That's is clean and easy to use. Most of the time it comes in those squeezy bottles, an advancement in usability that has been absolutely awesome. Sometimes you have to really shake a glass bottle. Here's a pro-tip: try gently knocking the far end of it to loosen the ketchup and enable air pressure to assist gravity. It's very rare that anybody really plans ahead by keeping the bottle on it's head. It's not something that is at the top of mind for everybody.

What comes out is generally the same.

It's been standardized.

The consequence is that everybody gets used to it. Yeah, yeah, I have to budget ten cents for every burger I sell for ketchup. Whatever. Nobody ever really eats the burger for the ketchup. Nobody ever really says 'wow, that was great ketchup'. Certainly, I believe that most people would miss the ketchup if it wasn't there. They'd say, 'wow, what an incredibly dry burger', but they wouldn't always be aware that there was no ketchup there. And that's the thing. Most people can afford ketchup. It's a condiment. A commodity condiment. It's mass market.

Most would complain if the taste of the ketchup was off. Everybody wants to standardize. And they'd be fine with getting the generic store brand so long as it tasted like the standard. Maybe the label or the bottle isn't as nice. But it's basically the same taste. But if it's off by just a little bit, well, that's an outrage. There should maximal transferability between ketchup brands, with no tolerance for deviation.

And that's okay. That's ketchup. You know what you're getting.

Mustard, on the other hand, comes in loads of varieties. There's French's Yellow Mustard. That's pretty common. But then there's Dijon. And Honey Dijon. And Gray Poupon. And a whole sequence of old world mustards like Tewkesbury horseradish mustard.

You never quite know what you're going to get, and it's frequently to taste once you get familiar with it. Lots of people relish the mustard. It completes the burger. The vast majority of people really don't know what they're missing. They got no clue. So, naturally, so many people go without the deliciousness of mustard to complete their experience. They never experience the tangy highs of prickly bear honey mustard. They never fully experience the vinegary flavor of russian mustard. And naturally, unless you naturally seek those out, and see the value in that, you're never going to really seek them out, and pay for them. Now, sure, you have to budget 20 cents for a really nice mustard for any burger. But it's well worth it. And makes the whole experience different.

And then there's the burger.

You'd notice if the burger wasn't there. The burger is the whole point. It's the whole reason for coming to the burger stand.

Burgers vary in quality and size. You got your standard McDonald's and Burger King Burgers. One is round and one is square. They're cut to the same standards and are the same anywhere. You have your Toronto Street Meat burger. It's pretty much the same everywhere. You have those Gourmet Burgers that tend to be far thicker and contain a better meat to fat ratio. There's a fair bit of variety in that too. And then you have destination burgers. Like, burgers made from Kobe beef, and just the right amount of fat added. They're also amazing because they can be cooked to medium-well, with a juicy flavor explosion so amazing and dazzling. It's like a party in your mouth and everybody is using a watergun.

Now that I've appeared in multiple social media monitoring accounts (hi guys), I'll unpack this whole analogy.

The industry is centered on ketchup because ketchup is predictable and comparable. Never mind that we don't really actually engage in real comparisons. But it's rooted there. And is it any wonder why nobody really respects the ketchup. It's just ketchup. It's alright condiment. I'd categorize a lot of the activity going on out there as ketchup. Sure, it's important and it complements a burger. But really. No.

There's a thriving industry around mustard out there. It's a niche industry. We got some people doing some pretty amazing things with usability analytics and morphing. We have an emerging neuro-analytical industry (a particularly spicy kind of mustard). And that aspect is very, very good. It costs a little more. But fundamentally, it goes with the burger.

The condiment industry has one thing in common with these types of analytics. Nobody ever sits down and eats a whole container of mustard or ketchup. [Edit: For the one person who emailed me telling me that he does, indeed, eat a whole container of mustard. Congratulations.] Mustard doesn't eat like a meal. Neither does ketchup.

And what of the burgers? Burgers certainly eat like a meal. They're the main attraction.

The fundamental challenge is how to get out of the mindset of forcing ketchup

--

Note: Special thanks to Stephane Hamel who provoked this rant at eMetrics Toronto.