Is Big Data Trumped? Lessons Learned for Marketing to Real People
Consider this post a fan-mail to you, association marketer—here's why we should not end up replaced by bots in the next four years.
Data makes the world go round in 2016. In our interconnected world filled with "big data," almost every decision is underpinned by numbers. That's why the spectacular failures of almost every pollster on Brexit and Trump's presidential win in the US has been so unsettling to marketers.
We have never had so much data available before. But, if everyone using big data can be wrong for the outcome of a national election, can we marketers trust data for comparatively small brands? Is the era of market research, personas, and all types of automation and optimization in jeopardy?
As a profession we have come far from the days of Mad Men (<--yes, obligatory cliche reference inside marketing blog post) and today's marketing seems all about optimization, ever-closer targeting, and proving ROI. In the relentless march toward efficiency and optimization, we've tried to measure every member action and response—tying ourselves as marketers to numeric KPIs. That can be a good thing—membership software can be extremely effective at measuring member engagement. The challenge comes with how do we as humans seek to interpret the data software gives us?
Data is only as good as its interpretation.
You will be a better marketer for understanding that. I'm sharing with you how I was reminded and what I think it means for us as marketers.
The polls (Brexit and Trump/Clinton) were wrong this year for two reasons:
- People are social beings who often make choices emotionally, not bound by the economic assumptions of Rational Choice Theory.
- The structural limitations of data generated by polling methodology cannot keep pace with consumer tech.
1. People Are Irrational Actors (i.e., People)
How many times do we write a blog post according to "data-backed best practices" and never pause to ask ourselves, "Would I actually read this in my own spare time?" Over-reliance on data can mean your marketing energy goes to the wrong places.
Chelsea Scholz hits this point hard on the Call to Action podcast episode "Behind Every Click is a Person."
We get really focused on doing things like writing emails to see click-throughs or optimizing a landing page so that Google recognizes it. And that really ends up making us forget the who that we’re actually sending these marketing initiatives to.
We’re sending emails to our customers and we’re creating customers from our leads. Those are people back there. And they’re ultimately what drives our bottom line; not the metric. So I find it to be kind of like one of those things that seems so obvious, it’s not obvious.
As a membership marketer at HSLDA, I was downcast on the days a new landing page didn't reach 2,000 visits. So as a B2B marketer for Aptify, how can I handle 175 people reading this post? Actually, a click means you came here to read what I have to say, and in a little way you're allowing me to affect your life and who you are.
As marketers, if we think of our work as life-impacting for actual real, live individuals, we will 1) be encouraged (whether in a low-traffic B2B environment with low metrics or in a B2C environment where the faceless mass of digital customers generates overwhelming amounts of data), and 2) create better marketing.
When thinking about your customers as people you're not going for the gimmicks and stupid stuff that gets you more clicks but ticks people off. You create more value for the customer and in the long-term have a sustainable brand that does not rely on a revolving door of customers. (You might be able to clickbait someone into joining, but they will only renew membership if you deliver personal value.)
In SEO, marketers used to face the question of writing for people or writing for search engines. Yet, today, search engines increasingly seek to punish metric-gaming and instead promote results for the needs of real people. Writing good content for people now gets you better search (as it should do).
Ask "What is a potential customer trying to do on my site?" and then optimize their journey. You will deliver far better topline revenue results than a fixation on A/B tests for button color.
If you think about join-engage-renew as a unified customer experience over time, you will better engage members than if you have 3 different departments looking at 3 different sets of KPIs independently.
So that is how the often "irrationality" of actual people affects how we should trust and interpret data. Modern technology itself is the second reason the polls were wrong and you need to interpret data critically.
2. Any Data Generated Has Structural Limitations
Cell phones enable area codes to follow voters around the country even as they move across state lines. Further, the generational differences in cell phone use and phone answering behavior are additional structural challenges in political polling.
Marketing faces structural problems with much data as well. For marketers, sometimes not every click is a person. Consider:
- Facebook admitted to five different calculation errors that overreported results sitewide for over two years.
- Bots online accounted for 48.5% of all web traffic in 2015.
- Bot nets distort trending topics and follower counts on Twitter (~50% of Clinton's and ~80% of Trump's Twitter activity was bot-automated).
So what is the response? Save the sanctity of data by changing your robots.txt on every webpage and placing CAPTCHA on every form or hyperlink?
Of course not, if you've read this post so far. You should never go so far in search of data that it destroys a user experience or customer journey.
Over-reliance on data without proper interpretation will stab you in the back though. Learn the context. Admit to yourself most web traffic and social media data has a wide margin of error and act accordingly.
- Place more emphasis on the data out of your membership software—from email marketing, events, and phone inquiries. It's less prone to bot manipulation.
- Make decisions and target messages based on data that is right-sized for you and your association. If hyper-targeting is beyond your skill, capability, or scope, don't guilt yourself into mediocre or bad use of big data. Even Proctor & Gamble has pulled back from hyper-targeted Facebook Ads.
- Big data has driven disruption, but encourage yourself and teammates to innovate by regularly stepping back to think outside the data box. The Clinton campaign didn't do that in Michigan and Wisconsin until all the ballots were cast.
Don't abandon big data to go back to hunches and HiPPO (highest paid person's opinion). Instead, remember, you only get as much insight from data as the effort you put in to probe and question it.
Here’s what several association influencers shared about data in their predictions for 2017.
Associations will get better at using data to learn more about their community and to deliver more value to them. ~ Deirdre Reid
I would like to see associations move away from a simplistic use of data to “segment” and “target” members and other audiences based on very broad and often inaccurate categories and begin treating them as real people with whom we have real relationships. ~ Elizabeth Engel
Read all the association influencer predictions here.
About Tony Cavicchi
As the Inbound Marketing Manager at Aptify, Tony delivers content aligned to answer organizations’ needs and show how an Aptify AMS will help them better achieve their missions. He knows many challenges faced by the association market firsthand after running digital marketing for three years at HSLDA, an Aptify client. Tony was a political science major, so he’s still a bit of a nerd and follows election news around the world like sports championships.