It is 2017 and our country’s politics are more polarized than ever.
It is 2017 and Donald Trump, the person my liberal friends dreaded the most, is now the president of the United States.
It is 2017 and over 80% of small and medium sized businesses (we have interviewed over 100) still struggle to understand what data can do for them.
You may ask: “What? What does the adoption of data technology have to do with political polarization?”
Although they may seem totally unrelated at first glance, in fact they share some important commonalities:
- They both reflect a failure to drive meaningful behavioral changes. Because neither adequately address the emotional drivers of behavior both the liberal community and the data science community have failed to drive meaningful behavioral changes in their relevant stakeholders.
- The field remains a confusing arena for “experts.” The election of Donald Trump came as a shock to most of the political “experts” and analysts . Likewise, a data scientist asked me the other day, “If AWS and Azure have made data analytics easier than ever then why are you still working on helping people adopt analytics?” Each situation reflects a fundamental misunderstanding of the situation.
- They are both urgent problems that need to be solved. In both cases, we should use human-centered design principles to better understand why people are not behaving in the ways “experts” expect. In the 2016 election, both the mainstream media and the left made many assumptions about how the American electorate will behave, many of which turned out to be false.
Although I am both a liberal and a data scientist, the purpose of this article is not to take a political stance or to criticize political ideology.
Rather, though this comparison I aim to better understand two behavioral puzzles confronting the liberal community and the data science community.
What the field of psychology says about behavior change
According to psychologists, there need to be at least three elements present in order to convince you to change your belief system despite your risk aversion bias:
· Low perception of risk: if the stranger is someone that you trust and have dealt with before, you are much more willing to make the change.
· Immediate reward or detriment: If the change causes immediate harm or benefit (providing 1 month of food instead of 1 day), then the push for change will be stronger.
· Low effort: if the effort of change is low, then change will also be easier.
If the situation does not satisfy any of the requirements above, risk aversion kicks in. No matter how good the long term benefits may be if you change, these elements are crucial in order to make the final push.
On the other hand, it also may be true that attempts to change behavior may actually have a negative effect on changing behavior. Simply put, we may cling even harder to old beliefs in the face of change because they are familiar and comfortable.
The situation described above is well-studied in psychology and termed “terror management theory”.
A classical experiment done in this area presents smokers with two smoking campaigns. Each showcases the detrimental effects of smoking (a blackened lung), but one includes a solution (some smoking assistance group or quick tip to stop smoking) and one does not.
The campaign without a solution intensifies the smoking behavior of participants, whereas the one with a solution decreases the smoking behavior of participants.
This experiment shows that without low effort and low risk actions to change, no matter how strong the immediate detriments are, people will nevertheless regress back to their old behavior, further illustrating the importance of all three factors in driving behavior change.
The left did not adequately apply human-centered design principles to persuade voters
Now let’s examine the messaging of the left from a risk aversion perspective.
By deciding to vote for a leftist or liberal candidate conservative voters have a lot to lose and very little to gain.
On the loss side, they must face the potential mockery and exclusion from their group (i.e. their conservative friends and family). This is one of the deepest fears in human psychology.
What it is there to gain? Almost nothing! To them, there is no immediately obvious or perceived benefit, emotional or economic, that would convince them to “convert” to liberalism.
Even if they were more moderate or open to liberalism, they often don’t know what the immediate action steps they should take are. Actions advocated by liberal groups such as “speak out and speak up” are too stressful for the average person to perform because by performing those actions conservatives risk being excluded from their core friend group. As the old saying goes, “the Right looks for converts, the Left looks for traitors.”
When facing this uncertain situation, along with extreme pressure from the media and liberals to change their viewpoint, conservatives idle. They see no clear benefit and don’t know what action steps to enact. Thus, the conservative voter feels alienated and becomes more attached to the ideological messaging that makes sense to them. This, in turn, further increases the polarization of politics in our country.
Trump, on the other hand, centered his message around immediate benefits to working class people. His populist platform, though vague, focused on economic nationalism, negotiating better trade deals through protectionist policies, job generation, and urban infrastructure projects, all of which represented immediate benefits to the American worker. He marketed himself as a strongman who would protect the interests of the American working class.
Data adoption faces similar challenges
Many of the problems associated with political polarization also apply to the question of data adoption in small and medium sized businesses.
What the data community needs to realize is that, even though we are data experts, most people in the country are not.
Data analytics became a buzz word around 6 years ago when McKinsey published an article stating both the important of big data analytics and the shortage of data analytics talents in the country (a gap of 140,000–190,000 analysts and 1.5m data-driven managers and analysts by 2018).
This means that small and medium sized businesses with limited access access to expensive technical talent have a difficult time understanding what data analytics really is and what it can do for them.
They, like the conservatives, see the data in their world and feel the need to be data-driven but none of the criteria for sustainable behavior change are satisfied.
In terms of risk, adopting data technology usually require at least a couple hundred dollars of investment per month for a business. They often also need additional consultant help to make sure the technology is working as intended.
Thus, due to the technical nature of analytical tools on the market there is no large immediate reward. Rather, it is very difficult to derive immediate business insights from data without the help of prohibitively expensive consultants.
Many data companies urge small and medium sized businesses to adopt a “growth mindset” when approaching data because it will provide them with long-term benefits. But that’s just not how our mind works.
Lastly, I already briefly mentioned the fact that data analytics is hard to adopt. Buying a tool is usually not enough. In order to fully take advantage of data, small businesses need to change their corporate strategy to be data-driven. This is usually at least a year long process.
Clearly, none of the pieces are falling into place. No wonder small and medium sized businesses are not adopting data analytics.
How to reduce barriers to behavior change and push people to change their behavior
The past is the past and there is nothing we can do to change it but the future can be better. In order to make it better we need to understand what we have done wrong both in preaching liberal ideology to voters and data science to small businesses.
Should we blame those people that became “racist” or “Trump supporters”? Should we blame the small businesses for being backwards and old fashioned regarding data?
No! Of course we should not blame human beings for what what comes naturally to them — the desire to be certain and secure.
What we should do is focus on what we can change. In order to do that we need to work with the conservative and data community to persuade them.
As a member of both the liberal and data communities I recognize that we cared too much about our concerns rather than working out what those we want to convince thought. Only when we work TOGETHER with them can we make change possible.
And because of that, we need to change our message.
If the left’s goal is to represent the interests of working class people, we should start with what their needs are and what emotional drivers resonate with them. One big theme from this election is that working class people feel like the political establishment on both sides only represent the interests of the elites. They felt alienatied and were deeply unhappy with the status quo of their country. We must start with their everyday reality, not the theory of the “experts.”
Honestly, creating messages that facilitate change isn’t the hard part. What’s really difficult is overcoming our emotions so that we can be consciously aware of other people’s feelings and perceptions regarding topics that people care so much about.
More specifically, based on what reduces uncertainty, we need to craft a message that:
1) Reduces the risk of change: put yourself in the shoes of your audience and think of their biggest risk in making the change you want them to make. Reduce those risks for them.
2) Increases the clarity of immediate value: instead of emphasizing the long-term benefits of the change, focus on what your core audience can get RIGHT NOW as a result of the change.
3) Decreases the effort required: make it easier and less emotionally taxing for the change to happen. Start with small changes and gradually ramp up the request.
I am not a political activism expert so I am not going to elaborate more on political activism solutions beyond this point.
However, I am very aware of the data challenges of small and medium sized businesses. We’ve talked to over 100 of them regarding their data challenges and know where they’re coming from (read our challenge analysis here).
At Humanlytics, the company I founded in an attempt to solve the data adoption challenge, we collected feedback from small businesses and used that information to craft the following solutions for the most challenging obstacles to technology adoption.
1) Low perception of risk: We build our solutions on top of companies’ existing data infrastructure (Google Analytics, for example) instead of being invasive and changing the company’s entire data structure.
2) High immediate reward or detriment: We have identified five core questions Google Analytics can answer for business owners and focus solely on helping them see immediate benefit in those four areas (Link to our questions here).
3) Low effort: Using Artificial Intelligence and human-centered design principles, we not only help businesses skip most of the burdensome analytical work (and required training or hiring) by providing automated recommendations, but we also organize analytical results in easy-to-understand business questions that help business owners quickly convert insights into actions.
It is our hope that, with our effort to reduce the uncertainty and pain points around data analytics, small and medium sized businesses can finally benefit from data analytics and improve the overall well-being of our society.
To stay updated on future blog posts, please follow us on Medium at analytics-for-humans, or on Twitter and Facebook. If you have any questions about this article, please feel free to email me at firstname.lastname@example.org.
This blog is produced by Humanlytics. At Humanlytics, we are making tools to make Data Analytics easy, compelling, and valuable for all businesses. If you want to learn more about Humanlytics, please visit our site at humanlytics.co.