In 2019, data drives everything — so much so that a Google search for “data drives everything” yields 169 million results. (“Data drives nothing” turns up only 107 million, in case you were wondering.)
That’s especially true in the business world, which creates a Catch-22 for professionals without statistical training. They can’t ignore the data their company collects on web traffic, sales, consumer behavior and more — often referred to by the umbrella term “business intelligence” — but confronting an unmediated database can be overwhelming. (Even Microsoft Excel gets flummoxed by the modern profusion of data, hence the rise of big data platforms.)
Business intelligence software offers a solution. It functions as a user-friendly skin over a database, translating its contents into digestible, actionable insights that people without data science degrees can grasp. Unsurprisingly, business intelligence software plays an essential role in organization-wide data cultures, which aren’t specific to tech anymore; at this point, even grocery chains cultivate data cultures.
San Francisco-based Heap is one of the many business intelligence providers. The company’s software tracks how consumers interact with digital products like websites and mobile apps. Some similar software only captures data upon request, meaning you have to write one applet to track how many people click a certain button, and another to track how many click a different button. Though functional, this can cause backups if clients realize, mid-test, that they need to track a different type of activity — say, zoom-out pinches.
That’s not the case with Heap, which by default collects every possible piece of data on how a user interacts with a product: clicks, taps, pinches, swipes — you name it. It stores all that data in a searchable, user-friendly Cloud repository. If an organization suddenly needs a new type of data, it’s immediately available.
Business intelligence tools like Heap’s have prompted a host of intriguing business decisions and strategies. With some help from the team at Heap, we rounded up interesting examples from five industries.
There would be no business intelligence without the tech sector. Companies couldn’t capture all this business data in the first place without Cloud computing — a concept that has been kicked around since the 1960s, it only became a real part of digital infrastructure in the 21st century. That’s thanks in large part to Salesforce, founded in 1999. One of the first SaaS companies, the renowned CRM provider played a key role in mainstreaming Cloud usage.
The tech industry didn’t just create the hardware for business intelligence, though. Tech companies themselves collect and analyze business intelligence in various ways. Fun fact: Even business intelligence software companies use business intelligence software.
Google: Data-Driven Performance Reviews
Location: Mountain View, Calif.
How it uses business intelligence: Google’s vigorous data culture encompasses a team within HR called “People Analytics.” It’s dedicated to bringing data to bear on employment-related questions, like how managers should get reviewed and, more broadly, whether managers should exist at all.
That’s right, one question they researched was “Do managers matter?” The answer, they found, was yes. At Google, anyway. Employees with better-reviewed managers performed better on the job and stayed longer with the company. The team next dug into what makes a great manager, drawing on interviews with the top- and bottom-rated managers at the company. They extracted eight keys to managerial success at Google, like “is results-oriented” and “does not micromanage.” Those findings, rooted in data analysis, became the foundation of their manager performance reviews.
Grow: Prompt Visualizations
Location: Lehi, Utah
How it uses business intelligence: This business intelligence company worked with Heap to unearth what users wanted from their free software trials: a data visualization. Though it was only a rough draft, it offered a sense of what Grow’s product could do. In fact, Heap’s data showed that users who created visualizations in their first session were five times more likely to become paying customers.
That came as a surprise. Grow initially had assumed that new customers preferred accuracy to speed, and their free software trial had started with a lengthy setup process. Users had to connect Grow to all their disparate data sources — their CRM, their web data, their finance data, etc. — to start creating visualizations. Heap’s data prompted Grow to streamline the path to visualization, though: now it takes just 15 minutes as opposed to 90, according to the team at Heap.
LendingClub: A/B Tests Galore
Location: San Francisco
How it uses business intelligence: This digital lender once struggled with A/B testing, a type of product optimization by which two slightly different offerings are tested on an audience to see which one they prefer. LendingClub’s team could rattle off myriad tests they wanted to run, but the testing process was so time intensive, they only ran about four per year.
But that changed when they became a Heap client. Heap’s maximalist data capture made testing simpler. The influx of business intelligence meant the company could suddenly run 12 A/B tests a week — more than 100 times their previous rate. LendingClub became the most frequent tester on Optimizely, a popular A/B testing platform, according to Heap. This led to hundreds of small tweaks to LendingClub’s digital storefront and reportedly boosted revenues by millions of dollars.
E-commerce has fundamentally transformed the retail sector. For one, the digital sphere makes it easier to capture data, so online retailers have an unprecedented amount of business intelligence. Digital native brands have also started opening brick-and-mortar stores, which means data from e-commerce now informs physical stores. Take Amazon bookstores: the displays feature star ratings from the web, and a “Quick Reads” endcap highlights books Kindle users read in three days or less.
Besides informing physical spaces, digital data also links with business intelligence from physical stores. Facebook, for instance, offers a tool for tracking how successfully digital ads prompt store visits. Increasingly, companies track and cater to shoppers’ purchasing habits both on- and offline.
Payless Shoes: Slowed-Down Checkout
Location: Topeka, Ka
How it uses business intelligence: Before shuttering U.S. operations earlier this year, Payless Shoes discovered something interesting about its customers: they didn’t want a one-click, ultra-streamlined buying experience. Data from Heap showed that Payless shoppers actually preferred a two-click process, which allowed them to review their carts before purchase.
This flew in the face of conventional wisdom, which holds that fewer clicks equals a better customer experience. That principle underpinned Amazon’s two-decade-old investment in one-click buying, which it immediately patented. But not every customer wants such a streamlined experience; it depends on the clientele.
Stitch Fix: Algorithmic Styling
Location: San Francisco
How it uses business intelligence: Stitch Fix offers clients a personal stylist that’s part human, part algorithm. The digital service works like this: Customers fill out a brief questionnaire about their style, and receive a personalized box of clothes in the mail. They pay only for items they keep and ship the rest back free of charge.
Stitch Fix collects business intelligence throughout the process, meaning the more a customer shops with Stitch Fix, the better the styling team (and their algorithm) grasps that person’s sartorial taste. In fact, the company hired multiple astrophysicists to decode the different layers and “notes” of personal style using a technique called eigenvector decomposition — complex work that would be impossible without business intelligence.
Target: Personalized Direct Mail Ads — But Not Too Personalized
Location: Minneapolis, Minn.
How it uses business intelligence: Back in the aughts, Target’s tracking software noticed a customer shopping in such a way that typically indicated pregnancy. (Think pregnancy tests plus maternity jeans.) And so Target automatically began sending that shopper coupons for cribs and diapers. It was standard operating procedure. There was just one problem: She was in high school.
“Are you trying to encourage her to get pregnant?” her irate dad said in a call to Target, according to the New York Times.
The manager he spoke to apologized profusely, but it turned out the man’s daughter was already pregnant. Target’s algorithms just happened to notice before her father did.
Target now structures its direct mail ads a bit differently in order to keep them from crossing the line between “useful” and “creepily useful.” According to Heap, half are personalized for the recipient and half are random, which helps customers discover new products without feeling surveilled.
Some pundits used to worry that free social media sites were overvalued. Now, though, it’s clear they serve a two-fold purpose: connecting and entertaining users, while collecting business intelligence that allows advertisers to run hyper-targeted campaigns — for a fee. (On Facebook, for instance, jewelers can market themselves specifically to recently-engaged people.) Social platforms also use their terabytes of business intelligence to perfect their own platforms — sometimes in surprising and high-tech ways.
Twitter: A Less Limiting Character Limit
Location: San Francisco
How it uses business intelligence: Back in 2017, this social media platform made the controversial announcement that it was doubling its 140-character limit on tweets; posts could now be up to 280 characters.
Plenty of users saw that as the death knell of everything good on Twitter. The platform would now host only wordy screeds, they predicted— no more pithy jokes and digestible anecdotes. But then… it didn’t happen. What Twitter had, and users lacked, was business intelligence. During testing Twitter had discovered that, on average, users with a 280-character limit posted tweets that weren’t much longer than before. They also spent less time adjusting for character limit, which under the 140 rule, often meant revising and condensing. In short, the change made Twitter’s platform easier to use.
Facebook: The End of Likes?
Location: Menlo Park, Calif.
How it uses business intelligence: Because Facebook is so widely-used and such an entrenched part of American culture, academia has begun collecting data on how the platform affects users. Though the conclusions aren’t always flattering, the data still functions as business intelligence.
Recent studies, for example, have shown that Facebook use correlates with a dip (though perhaps a small one) in users’ mental health. One potential root cause: users’ ability to compare their actual lives with other people’s carefully-curated representations of their lives. This apples-and-oranges comparison leaves nearly everyone feeling lacking; the option to compare “like” counts only compounds the issue.
Perhaps inspired by this research, Facebook recently began experimenting with making like counts private so that only the original poster can see them. While the test isn’t yet live in the U.S., Facebook is investigating how it performs in Canada, Japan and a handful of other countries. The exact motives behind it aren’t clear, but tech experts have hypothesized that it’s an attempt to reduce envy.
TikTok: No Need to Browse
Location: Los Angeles
How it uses business intelligence: TikTok leverages user data so savvily that it has made the act of browsing near-obsolete. Simply download the app and open its “For You” tab and it immediately starts playing entertaining short-form videos. By reacting to the videos, users teach the app’s algorithm their likes and dislikes. Hard-pressing on a video indicates “not my jam,” whereas a heart means “I like this.” Over time, users hard-press less and less, and the company refines its algorithm more and more — or so the theory goes.
As Jia Tolentino wrote in the New Yorker: “Some social algorithms are like bossy waiters: they solicit your preferences and then recommend a menu. TikTok orders you dinner by watching you look at food.” Though the app’s advanced technical capabilities have the potential for misuse, its algorithm is also a case study in artfully leveraging business intelligence. As Tolentino noted, “I often found myself barking with laughter, in thrall to the unhinged cadences of the app.”
Food & Drink
Tracking the American diet can be a bizarre business. We eat too many calories, but we also have trouble finding the time and energy to eat dinner?
Still, that’s exactly what companies in the food and drink sector do. They collect business intelligence on who eats what, at what hours, during which seasons. It allows for personalization, and keeps our favorite restaurants from running out of chicken tikka masala during rush hour.
Starbucks: An App for Regulars
Location: Seattle, Wash.
How it uses business intelligence: Starbucks’ popular rewards app was used by more than 23 million people in 2018. Although it offers consumers rewards, like free drinks, its popularity is also a data boon to Starbucks. The coffee chain uses information it gleans about individual purchasing habits to target customers with appealing deals, so they repeatedly return to buy more of their favorite lattes or muffins. And while it’s true this strategy works only for customers with smartphones, the app isn’t a new customer recruitment tool; it’s more about retaining regulars.
Kroger: Custom Coupons
Location: Cincinnati, Ohio
How it uses business intelligence: Like many chain stores, Kroger tracks each customer’s purchasing behavior through a loyalty program. Unlike many chains stores, though, Kroger has a wildly successful direct mail program — more than 71 percent of the households it mails to redeem at least one coupon in-store per quarterly mailing, Forbes reports.
The tracking data that informs the mailing is robust enough that Kroger can avoid marketing to customers in certain segments — instead, they can market to individuals, customizing the layout of their mailings to each and every consumer. Eighty percent of the dozen coupons in each mailer apply to products the recipient demonstrably likes, the remaining 20 percent to products they might like based on what they’ve already purchased. The approach has so far brought in revenues of more than $10 billion.
Uber: Business Intelligence for Restaurants
Location: San Francisco
How it uses business intelligence: In 2015, when Uber launched its food-delivery app UberEATS, Uber’s interests became entangled with the interests of thousands of restaurants across the country. The mechanics of the UberEATS app also meant that Uber constantly collected business intelligence on those restaurants — which was interesting, but not particularly actionable for Uber.
So in 2016, Uber’s engineers built Restaurant Manager, a scalable, user-friendly analytics tool that gave their partner restaurants access to real-time metrics like net revenue, number of daily and weekly items sold through the app, order prep speed and basic customer satisfaction (eg., Did they give their meal a thumbs-up or a thumbs-down?) Overall, the product gave restaurant owners a clearer sense of how specific dishes tasted on certain nights.
The transportation industry’s business intelligence has an Achilles’ heel: It can’t solve the traveling salesman problem (or, as insiders call it, TSP). This logistical problem has plagued logistics professionals since before “logistics professionals” were a thing.
The basic problem is this: If a traveling salesman needs to go to 100 different locations over the course of a day, what’s the best route for them to take? It seems simple enough, except for the fact that no one knows. There are just too many possible routes — even on modern computers, algorithms that could solve the problem work at impractically slow speeds. That means certain aspects of routing and delivery models may never be perfectly optimized. Still, business intelligence has improved the travel and transportation industry in other ways.
Navistar: Predictive Analytics, Meet Truck Repairs
Location: Warrenville, Ill.
How it uses business intelligence: This transportation company doesn’t just manufacture trucks, buses and military vehicles, it also makes fleet management systems like OnCommand Connection. The Hadoop-based software keeps trucks connected to their home base while they’re on the road, capturing 100 daily data points per truck and relaying them back to home base. That helps Navistar and a host of clients predict when trucks will need repairs, which reduces unnecessary maintenance and unexpected breakdowns on lonely highways. In business terms, it reduces downtime and boosts return on investment for vehicles.
Delta Airlines: A Baggage Tracking App
Location: Atlanta, Ga.
How it uses business intelligence: The hold of an airplane is a black box, literally and figuratively. What goes on with the luggage in there? Well, Delta tracks that with an elaborate system of bar-coded baggage tags and sensors. The airline shares this location data with its customers through an app, to demystify the baggage handling process. It’s pretty popular, too. Currently part of the Fly Delta app, the baggage tracker was originally a standalone app that got more than 11 million downloads.
Tesla: Business Intelligence as Brand Defense
Location: Palo Alto, Calif.
How it uses business intelligence: Tesla’s remarkably autonomous cars are all Internet of Things devices that are digitally linked to Tesla HQ. The setup allows Tesla to push software updates and new digital features to old cars, free of charge. It also lets the company collect business intelligence on how Tesla owners drive their vehicles.
That data comes in handy, and not just for patching glitches in real time. Back in 2013, the Tesla’s networked nature helped the company push back on a New York Times review of the Tesla Model. The reporter alleged that on his test drive the vehicle’s heat didn’t work, and that the car struggled to achieve 45 mph on a low battery. But according to Tesla’s data on the vehicle, the heat actually did work — the journalist just accidentally turned it down instead of up — and the car drove faster than 45 mph for the entire ride. Ultimately, the Times’ public editor concluded that the test drive was undertaken in good faith, but the journalist “left himself open to valid criticism by taking what seem to be casual and imprecise notes along the journey, unaware that his every move was being monitored.” In other words, he didn’t intentionally lie — but neither did Tesla’s data.
Images via Shutterstock, social media and company websites.