From Facebook’s $5 billion fine for mishandling user data to Amazon’s botched plans for its NYC headquarters, consumer trust in big tech nosedived in 2019. Negative views of tech’s impact on society doubled between 2015 and 2019. Likewise, the number of people who believe tech has a positive impact on the U.S. dropped from 71 percent in 2015 to 50 percent in 2019.
Trust in tech is especially low among some of the most influential subsegments, like millennial and multicultural consumers. But believe it or not, there’s good news for small businesses.
Research from Gartner (available for clients) shows that U.S. consumers have high degrees of trust in locally owned businesses and small brands. With 2018 trust levels at 90 percent and 87 percent (respectively), people trust them more than their communities (72 percent), corporate America (27 percent), and the government (19 percent).
It’s worth noting that in many sectors where trust declined, it was low to start with. For example, U.S. consumer trust in the government held at 36 percent in 2016, before dropping to 19 percent two years later.
If you’re a small business owner, you have the advantage of high trust levels now. To keep it that way, you must show consumers that your business handles data in ways that are as fair, accurate, and transparent as possible.
Since data management is a sticking point for consumers—and 2020 started with more regulation—this is the best way to keep your consumers’ trust. The following business intelligence (BI) trends can help you reach this goal:
Builds trust by: Informing and improving the decisions leaders make.
65 percent of global enterprises plan to increase their analytics spend in 2020. That’s no surprise to Gartner, where questions about BI and artificial intelligence (AI) are top concerns for business and IT leaders alike. It’s clear that analytics ownership needs to expand, and 2020 will be the year it starts in earnest.
Historically, data and analytics have been owned by one person in an organization. The most senior technical leader (normally the CTO) often makes most, if not all, big decisions about analytics use, from which tools the business will use to which data employees should access.
This approach creates several problems. When data sits within just one leader’s purview, that person often struggles to share the full range of benefits that data and analytics can offer the business. This leads to detachment on behalf of other team leads, who can’t access these tools to solve their own business problems.
Likewise, BI strategies that come from technical teams tend to be limited in scope and too focused on technology. Without addressing the unique challenges that leaders in different lines of business face, even the best BI software will only take teams so far.
In 2020, leaders in all lines of business will start using analytics. This will be due in part to shortages of data scientists, who are in high demand but short supply.
But the main reason is due to broader business needs. Whether one works in IT or marketing, the volume of data available today means that no line of business can afford not to use it. That means a wide range of leaders will need BI and analytics tools to help them do so.
Builds trust by: Empowering data scientists to boost business outcomes.
Despite high demand for data scientists, many are dissatisfied in their current roles. A survey of 907 UK data scientists in Autumn 2019 found that 56 percent said they plan to seek new roles in 2020. When asked why, 44 percent blamed bureaucracy and 29 percent cited lack of support from leadership.
This is in part due to the disconnect between the skills that employers demand of data scientists and their lived experiences at work. A 2016 survey found that 60 percent of respondents said they spend most of their time cleaning and organizing data. In the same survey, 57 percent said these tasks are the least enjoyable part of their work.
That’s due to change in 2020. Gartner predicts that by this year’s end, more than 40 percent of data science tasks will be automated. As a result, data scientists will see their roles change drastically this year.
Instead of cleaning and organizing spreadsheets, data scientists will spend more time building new models. Driven by the desire to apply data everywhere, business leaders will give their data scientists the power to monitor, integrate, and customize machine-assisted models.
That said, businesses risk minimizing their data scientists if they lack succinct strategies for how to use data. Without end goals for practicing data science, the failure rate for such projects will remain high: Gartner predicts that through 2022, only 20 percent of analytics insights will yield business outcomes.
So, if your business has no strategy for how to use data, creating one should be your top priority. Businesses that don’t know their “Why?” for using data (such as increasing consumer trust) will continue to see their projects fail in 2020.
Builds trust by: Offering transparency into algorithms’ decisions.
AI’s promise and value are well-documented, often to the point of overhype. It’s true that in some use cases, today’s AI technology is more accurate than humans at performing rules-based, repetitive tasks. As one example, some technology can diagnose patients with specific illnesses faster than physicians.
However, this does not mean that AI will replace doctors. On the contrary, Gartner predicts that AI will become a positive net job motivator in 2020, creating 2.3 million jobs while eliminating 1.8 million.
That said, with great growth comes great responsibility. The AI sector has a habit of training its technologies on biased datasets. This is due in part to more widespread use of black box algorithms, which prevents people from seeing how data points within a dataset interact with each other in the algorithmic training process.
In 2020, consumer demand for explainable AI will force businesses building and using algorithms to peek under their creations’ hoods. Consumer awareness of bias in AI is growing.
That’s attracting more scrutiny from governments and attorneys alike, which can result in lawsuits. A firm like Goldman Sachs can afford multi-million dollar fines; the same penalty will put the average SMB out of business.
So, as you plan ahead for 2020, you’ll need a more thoughtful approach to using algorithms. Design them to be interpretable from the start, rather than trying to interpret them after the fact.
Weigh the trade-offs of your desired business outcomes against regulatory requirements and customer expectations. This is especially crucial if you work in sectors that build products which make decisions about people, such as HR, healthcare, and financial services.
If the risk of building a tool that discriminates against customers exists, you’ll need to include product and platform evaluation requirements — as well as process oversight — into model algorithm selections and training data. Open source toolkits like Microsoft InterpretML and IBM AI Fairness 360 are an ideal first step to get started.