Marketing for manufacturers, stereotypes, data and hubris

Ed Marsh | May 16, 2017

what's the role for data and metrics in marketing for manufacturers.jpg

B2B Marketing stereotypes

There are many stereotypes in the world of B2B marketing for manufacturers. Two are relevant here.

  1. The naive soul who develops a project brief to "create a viral X" - could be video, infographic, article, tweet, etc. Everyone in the know guffaws at the idea that viral could be somehow engineered and deliberately created.
  2. The Warby Parker bespectacled, bearded data marketing savant to whom all genuflect in admiration for his (or in the unbearded version, her) ability to identify precisely why some marketing worked brilliantly, why other didn't, and how to replicate the success of the former.

Yes, I have chuckled at the former and lined up to learn from the latter.

Now I'm not so sure. 

I just finished reading The Drunkard’s Walk: How Randomness Rules Our Lives and it's challenged some of what I've accepted as fact in the world - including in the discipline of marketing for manufacturers.

Leonard Mlodinow schools the mathematicians too

Read the book. It will shape the way you advise young people on their careers, the way to plan your priorities and perhaps even how you manage relationships.

There is one example after another of randomness, and its cousin probability, that seem so counter-intuitive as to be outlandish. (Of course as a RedSox fan I have to disagree with the assertion that to consistently crown the better team, a baseball playoff series would have to involve more than 300 games. The '04 ALCS wasn't a fluke!!)

One of the consistent themes is that these problems are complex enough that even "experts" including mathematicians, statisticians and those whose professional judgment is influenced by probability (physicians considering treatment regimens) can be tripped up.

That got me thinking that if they can, then certainly each of us that are adherents of data based digital marketing methodologies could as well. 

What does it mean to marketing - theory

We ascribe enormous power to metrics and data in today's digital world of marketing for manufacturers. Web traffic, CTA click rates, landing page conversion rates, nurturing workflow goal rates, and email open & click rates are examples of metrics which we scrutinize for improvement opportunities. (CTRs on keywords is another that I wrote about last week.)

The question therefore is whether we imbue those statistics, which Mlodinow demonstrates to be far more random than we assume, with excessive importance.

The answer almost certainly, based on his book, seems to be yes.

If that's the case, then why are some marketers so consistently effective? He highlights examples from the investment world which indicate that track records are much more random than we'd like to think. So it's possible that marketing success is more random than we'd like to acknowledge.

It's also clear that there's a halo effect - when someone's had success, and/or is recognized as an authority, each of their insights is afforded more importance than it would have been otherwise.

But what about a landing page with 1,000 visits and a 23% conversion rate (vs. an average of 45%)? Or a workflow with a 3rd email that only unsubscribes at a fraction of others and results in far more clicks? Is there enough there for us to identify the reason and replicate the positive factors elsewhere?

Mlodinow indicates that our attempts to do so are often simply a process of identifying otherwise random precursors and ascribing correlation and causation in retrospect.

Boil it all down and the inevitable conclusion is disquieting - we may have less control over our tactical successes than we'd like to think.

And to marketing for manufacturers - in practice

We all know that a solid batter only gets a hit 25% of the time.

Content marketing guru Sujan Patel doesn't aim that high.

I expect – and plan – for most of my content to fail. On average, I expect just one in five of my articles to succeed

So the biggest takeaway may be to be more realistic in planning and anticipating results.

But I think there's a bigger piece here.

That's consistently reducing opportunities for randomness.

What do I mean by that?

Maybe it's unrealistic to conclude that a landing page on your site that converts at 50% is in fact better than one that converts at 11%. To identify key attributes that have delivered better results may be folly, since the delta between those results could be entirely random.

HOWEVER, there are certain things that we know will reduce variables which will impact randomness.

For instance:

  • write about business topics of interest to your target persona
  • use pithy email subject lines to get attention and focus on email deliverability
  • include details (section headers, bullet lists, images) that make it easy for readers to scan and engage
  • be relentlessly empathetic with buyers perspectives on issues of import to them as determined by qualitative research - not the stuff you assume is important to them
  • use multiple channels to boost distribution and engagement with your content to help more of the right people find it
  • clearly define your ideal prospect/customer profile and be pragmatic in pursuing it as Bob Apollo outlines here

Will we continue to look at 10% differences on landing page conversion rates for pages that we arbitrarily decide have a statistically relevant number of views? Yes. That's human nature; we're conditioned to believe that data is sacrosanct.

But maybe it would be healthy to temper that view with a bit of randomness reality....it would be less hubristic and perhaps more realistic. Maybe we're all a bit closer to the naive rube who orders up a viral video than we'd like to admit.

In the meantime, when it comes to driving deals across the finish line and creating revenue there are a lot of steps in a successful sales process. Have you ever looked at ways to automate some of those?

Marketing automation offers a powerful tool set beyond the routine download management for which it's often used. This free guide lays out more detail.

New Call-to-action