When Kinsa-founder Inder Singh, MBA ’06, HST '07, had an idea for a smart thermometer, he envisioned a device that could not only take temperatures but also enable users to monitor symptoms and medications, as well as keep records for their caregivers.
Work smart with our Thinking Forward newsletterInsights from MIT experts, delivered every Tuesday morning.Email Address
But here’s what really sets Kinsa apart: The San Francisco health technology company uses aggregated, anonymous data from its FDA-approved smart devices to track fevers and map where illness outbreaks are occurring in real time.
As Kinsa and myriad other startups know, the crucial component is data and analytics.
Startups find success when they possess information unique to their industry and apply analytics to interpret and deploy that data in strategic ways. Think of analytics as a key to an undiscovered kingdom: Wielded properly, it can unlock new worlds.
Consider these MIT Sloan-led startups that have analytics baked into their core business propositions:
Newton, Massachusetts-based Onduo focuses on Type 2 diabetes management, offering wearable glucose monitors to help customers monitor levels without a painful finger prick and delivering ongoing data to doctors and users. All member data is completely HIPAA-compliant; in the COVID-19 era, this kind of service — reducing the need for in-person doctor visits — is even more important.
Boston-based fintech Flywire streamlines international payments, partnering with schools, hospitals, and businesses to offer rates below those offered by many wire services and banks. Data from the billions of dollars in payments it processes allows Flywire to deliver value to its customers, maintain a competitive advantage, and achieve operational efficiencies over the long term.
And Kinsa can effectively predict where and when flu, colds, and now COVID-19 are spreading, connecting more than 1 million users with an AI-based medical triage system accessed through an app. Users can see which diseases are circulating in their area and plan accordingly. According to a recent New York Times story, Kinsa’s interactive maps have precisely predicted the spread of flu nationwide roughly two weeks before the CDC.
All businesses can learn from what these startups discovered at the outset — analytics matter. Here are some of their top lessons and tips:
Good data has a personal impact
Onduo’s product delivers immediate data and insights to customers: When they eat, they see how their glucose changes right away. “The learning is very obvious and very apparent,” said Gino Korolev, MBA ‘15, vice president of solution at Onduo, explaining that customers have instant incentives to modify behavior if they see a spike. “They learn they shouldn’t have sweetened cereal or that they should go for a walk after lunch,” he said, a lesson borne out by visible data instead of an abstract lesson.
Good data targets a niche market
Onduo spotted an untended niche, opting to target patients with Type 2 diabetes, also called adult-onset diabetes, rather than Type 1. “This was a novelty,” Korolev said, noting that the continuous glucose monitor industry typically focused on the “life or death” glucose level management of Type 1 diabetics. Now, Type 2 patients have access to a continuous glucose monitor that delivers constant data to doctors over an ongoing period, allowing doctors to precisely tailor medications for each patient.
Thrive, a San Francisco-based fintech, likewise identified an underserved market — graduating students in need of short-term cash as they launch their careers. Founded by Twitter alumni Deepak Rao and Siddharth Batra, who came to the U.S. as students and didn’t have access to traditional financial services, the firm’s ThriveCash offers underfunded students access to grants of up to $25,000 based on internship and job offer letters.
As the coronavirus crisis developed, the company was able to tap its data on offer letters and internships from hiring companies to quickly launch its Student Hiring Tracker, which tracks which offers are still being made and what programs are being canceled.
Data and analytics should be a top-level function
In 2019, Onduo hired a head of data analytics, who oversees five full-time employees. “About one year in, it was decided that data and analytics needed to be a separate function, a top-level function that will provide value to end users, our buyers, and the business,” said Korolev. “I think many different functions can act as analytics-capable functions — for example, engineering usually knows how to manage and query data. But it’s harder for an engineer to be in front of the clients, talking about how a program has certain efficacy for this client.”
The analytics team synthesizes data from clients and health plans, as well as runs projections, allowing Onduo to properly scale its business. The analytics team can evaluate past performance, develop projections for user growth, and give business guidance on onboarding new staff such as coaches and doctors.
Common objects make effective data-collection tools
Convenience increases adherence, which in turn begets more and better data. “Our thermometer is just a smarter, better version of something you already use. We piggybacked off of an existing behavior,” said Kinsa founder Singh. In this way, Kinsa didn’t have to convert users to a behavior to collect data — just enhance what was already there.
Impactful data solves a problem and spurs action
Kinsa distinguishes itself by aggregating data from customers and mapping it in real time, allowing sick (and possibly scared) users to get up-to-the-minute information to make decisions quickly: about whether to seek medical help, whether to go to the doctor, whether to stay inside.
“When we started Kinsa, the whole idea was: How do you get data on where and when symptoms are starting? How do you get data on how fast [something] is spreading? How do you get data on how severe it's going to get, and how do you get data from within hours of symptom onset — not days like the health care system gets,” Singh said. The immediacy is a differentiator for customers.
Strategically, Kinsa knew that simply providing data to users about where outbreaks were occurring wasn’t enough. They also needed to tell users what to do with that information, which happens via the app. “What we’ve built isn’t just a connected thermometer; it’s a connected medical guidance system,” Singh said.
Two minutes after users take a temperature, they receive a push notification asking for more information about symptoms; with each tap, questions become more specific, offering symptom-based guidance about what’s going on in the community and what to do next.
“This is the foundation of why we started the company. It’s to get better data, to analyze that data, so we can solve a problem,” Singh said.
Data should be shared widely across teams
Flywire’s analytics team relies on big-data analytics platform Looker, a company that was recently bought by Google, to share customized information across teams every day. Information isn’t siloed.
“Everybody in the company can access and check business performance and any of the metrics that they're tracking. So, in a traditional company where you would need an analytics person to create a report and present it every month, our business stakeholders create it themselves with support from analytics team,” said Mohit Kansal, MBA ’14, vice president of global payments.
“Anyone in the company can easily go and click to access these reports. It’s proved to be super useful at the company across all levels, not just senior leadership,” Kansal said. “In current times, where sudden business changes are very common across industries, it's critical to empower key stakeholders with the ability to quickly and easily monitor trends and adapt to changes in real time.”
Bring in analytics expertise early
“I think one of the mistakes that companies make is they run tests, and different teams run those tests, and when they're done, they might go to an analytics team and say, ‘Hey, can you help me analyze this?’” Kansal said. “A lot of times the answer is, ‘I can, but it won’t lead to any useful impact. You should have structured the test differently, which would have helped us measure it better.’”
Kansal always tells his team to bring in analytics from the outset and to partner with data experts to set up any experiment with a strong business context.
“Those are our best moments: When an analytics team understands our team’s work, they might come back and say, ‘Hey, I found something useful.’ The team will give you an insight that you’ve never had, and it has a significant business impact.”
For example, Flywire’s analytics team used machine learning to help design an automated matching system and reconciliation system for invoices and incoming payments, something that used to be done manually, because they were able to understand the pain point.
When analytics expertise is baked in from the beginning, it helps startups improve their products and grow in a more strategic way. “We’re seeing tech automation, we are using smart analytical tools, and that’s really one of the big areas that the team has been investing their time in, helping us scale this business. We might have ten times the size of the business — but we don’t need 10 times the number of people, because of the product and analytics team’s tools.”
In short: Don’t underestimate your data and analytics team.
Data and analytics, Kansal said, are “the source of truth and a big lever for growth.”
Source: https://mitsloan.mit.edu/ideas-made-to-matter/why-analytics-secret-sauce-startup-success