Hennepin County officials are finally deploying a data analytics system businesses have used for years to understand how customer behavior impacts the bottom line.
It's an approach typically seen in the health care industry, where providers track a patient's journey from check-in to discharge to determine how efficiently hospitals and clinics are functioning, which not only improves patient care but also reduces costly errors. Retailers use it, too. They analyze data on shopping habits to decide which promotions to run and when, which marketing campaigns are most effective and how to adjust return policies.
It wasn't until the county's top data scientist listened to a presentation from an executive at Optum — a health care services provider — a few years ago that the concept seemed applicable to the public sector as well.
For the past few years, the county's data science team has been developing a customer journey mapping system that will summarize — in a secure and privacy-forward way — the movement of citizens through the county's seven lines of business: disparity reduction; health; human services; law, safety and justice; operations; public works; and resident services.
The county's mapping system will likely launch later this year.
In the long-term, county officials hope the ability to make data-driven decisions will greatly improve quality of life for residents. In the short-term, it will inform decisions behind budget adjustments to expand, or wind down, various programs, Hennepin County Administrator David Hough said.
"We collect money from our residents, and in order to justify the use of those dollars, a lot of people expect you're going to be able to present data and show why are you doing this," he said.
The old way
Annually, hundreds of millions of dollars funnel into programs — designed to provide transportation, housing and food assistance, health services and employment training, and more — for Hennepin County's 1.26 million residents.
For decades, though, assessing the success of those programs has mostly come from department heads leaning on feedback from individuals and families. Every few years, third-party consulting firms conduct studies, digging into years-old statistics and feedback to provide insight to county leaders as to how to adjust or modify programs.
"We have a lot of data about our programs and the administration of those, but when it gets to the outcome of those [programs], it's actually a lot harder for us to identify," said Erik Erickson, Hennepin County's chief data officer. "We don't have employment [data] for every person. We don't have the income. We don't have the educational attainment. So for us trying to evaluate what programs are working and helping people get to a good place, those are data we don't have."
That's because county data systems are really fragmented, Erickson said. County departments developed programs in isolation and store data differently.
For some departments, data is at the case level, others by household, family or individual, Erickson said.
To deposit all of this into one visualization, data scientists are using a graph database model, allowing them to cluster data from different departments into a network of connecting points representing a resident's journey.
During the pandemic, county data teams used graph data to observe case volume and vaccine distribution. They expanded to other data sources, like anonymous cellphone use from Google, to gain insight into how county residents were moving around as COVID-19 cases decreased.
Applying data insight to every county line of business seemed like the next logical step.
"Right now, we're having to operate on a lot of anecdote and reports from staff about what's working, which isn't the same as an actual pattern," Erickson said. " We can't build an entire system around a compelling anecdote."
Data source
When a person schedules an appointment to receive a county-run service, that interaction becomes a record in a department database. The county updated those files to indicate a person's movement through the system, from a case closing to a program coming to completion.
The county can then analyze these numeric identifiers, not individual names, for patterns to determine gaps in service, Erickson said.
For example, resident A could have recently had a medical visit at Hennepin County Medical Center. Linking nodes from A would include the individual's insurance information, medical visit history, past diagnoses, procedures, treatment and medication.
By aggregating the data into summarizations, the county can provide insight to department leaders while keeping people's identity private and secure, Erickson said.
This data can also visualize churn, meaning how often residents enroll or withdraw from government programs, which can have both negative and positive connotations, Erickson said.
A person phasing out of a government program "is actually good," Erickson said, as it represents someone has used it to stabilize their life. If one were to re-enroll in pathways designed to remove them from a homeless situation or find employment, though, the county could then track gaps in service that failed that person and enact corrective actions.
"They're not getting the intended impact of that program to create self-sufficiency because they have to come back to us," he said.
A proven strategy
Analyzing how people interact with services using a mapping concept has worked successfully in the private sector for many years.
Oftentimes, company and organization leaders excel in operations but don't know what it's like to live the customer journey, said Jim Tincher, founder of Minneapolis-based Heart of the Customer, a customer experience and journey mapping consulting firm. He has worked in the customer experience industry for more than a decade.
A journey mapping system gives those leaders "perspective to help staff know what it's like to be a customer or a resident," Tincher said.
Tincher's clients through the years have included manufacturers, nonprofits and health insurance companies. Each client used the data-driven mapping system to change programming and strategy based on customer insight, which ultimately improved their profits or better satisfied their goals.
One client in particular, Dow Inc., a chemical company in Michigan, had an issue resolving complaints. Using a mapping system, they were able to decrease resolve time from 30 days or more to within day and a half, greatly reducing labor costs, Tincher said.
"It really helps organizations better understand and build empathy and drive change," he said.
For staff members, visualizing sets of data also helps communicate the message to those in power that it's in an organization or company's best interest to adjust strategies based on facts rather than assumptions.
First foray
The mapping system will be the first of its kind for the county, Hough said. And it might even cut down the amount of time people spend at a county service center for a new driver's license or car registration tags, Hough said.
"Where do we need to put the resources so that we're making sure that person going to Midtown Exchange Service Center has as short of wait as a person going to South Hennepin County Service Center?" he said. "All those things are really important so that we're delivering customer services in an efficient way."
Roughly three years ago, Hough began meeting regularly with Erickson to learn how the county could become more data-informed. Those conversations marked the genesis of the county's first data analytics department.
It's a two-fold benefit to a county looking to decrease the number of people using government assistance while repurposing those freed-up dollars to help those still in need.
"We're doing what I think government should be doing," Hough said.
