CDAO Alan Jacobson on hurricanes, data science, and accountants with "crayons"
Lots of companies "over-index on the most advanced thing for the smallest number of people"
Alan Jacobson, former director of global analytics at Ford Motor Company and now Chief Data and Analytics Officer (CDAO) at Alteryx, knows the power of data to help drive organizational change. During his career at the automotive giant he held a variety of leadership roles across engineering, marketing, sales and new business development – working to deliver analytics to the manufacturing plant floor, using data to optimise shipping and logistics, launching new forecasting tools within the finance organisation and more. As a result, he knows through hard-won experience that data democracies trump data dictatorships every time.
As he puts it bluntly: “I so often see huge money being spent on some of the most advanced analytics in the world, for a very small portion of their knowledge workers.
“The public sector can be a good example: the elite cyber-something-defence-force, have the billion dollar investment in AI and machine learning, but the accountants that are working for the government are still working with crayons, practically. You get that in different boardrooms too: everyone is ‘supporting’ AI and ML’ they’re all ‘supporting’ analytics and automation.
See also: Sol Rashidi, Chief Analytics Officer, Estée Lauder, on messy work, data culture, and playing pro rugby.
“But what I see is that some of them have very much over-indexed on the most advanced thing for the smallest number of people, to try to deliver some huge thing in a small pocket; versus widespread analytic maturity where everyone gets 10%, 20%, 30% better…"
He adds: "I do think more and more companies now are figuring out they need widespread analytic maturity: you should be able to walk up to any desk, ask a question and get a data-driven answer. Not write a ticket to the 'ivory tower of PhD data scientists' somewhere to get you a fancy answer. That's not analytic maturity. analytic maturity is we all have data literacy, we can all understand how to do these things to some level.”
It’s an unexpectedly passionate soliloquoy, and clearly a heart-felt one. As CDAO at Alteryx, Jacobson is keen to evangelise the company’s solutions, across predictive analytics, data mining, business intelligence, visualisation and data discovery, to name a few, but as a Changemaker, it’s the ability to empower people with tools that enable them to drive change themselves that really resonates with him again: that democratising power of digestible, data-driven insight that everyone has access to: “Helping people on that journey is very rewarding”, he notes, admitting wryly “I’m also very competitive! I like building the better mousetrap, as it were. That's why I went into engineering. I like solving problems. And data science provides wonderful tools to try to solve some huge problems. That certainly also gets me out of bed in the morning.”
What about bringing people along? As we’ve seen across the Changemakers series, that ability to involve people in owning disruptive change and -- also sometimes knowing when all the charm, force, or capability in the world won’t make an organisation ready for it – is crucial to what makes Changemakers tick. How does Alan ensure everyone’s on board when it comes to increasing the analytical maturity of an organisation? He admits its not always easy.
“You do get some people who think ‘we can't give analytics to everyone. They'll make terrible mistakes, they're going to destroy our business, because I'm going to give them AI and ML and they're going to come to completely the wrong conclusion and drive our business into the ground. And it is like being scared by the arrival of the calculator.”
“Imagine not giving accountants calculators because they don’t know how the technology of the calculator works; that we can't give them these modern day tools that do math, because we're worried about them making mistakes? The irony of it is it's often the IT teams and sometimes data science teams that are saying this: that we can't give the domain expert like the tax expert, this tool, because they'll make a mistake. Ironically, when the data scientist builds a model for for the tax person, it’s the tax person who’s going to be able to check it best.”
See The Stack's exclusive interview with JPMorgan's Global CIO Lori Beer
Ultimately, as he points out. “it's always a change management problem; getting people in on the journey, and some obviously sometimes struggle with that. I spend most of my time actually not teaching customers how to use the tool or the technology, I am teaching them about how to help them go on this change management journey.”
Are there any common errors he sees in driving data-led transformation? “If I were to list out the top 10 most transformational projects that I've worked on in my entire career, I’d that 9 out of 10, we would all look at and go, ‘that wasn't very Big Data, it was pretty Small Data’. So people do over-index on the wrong things sometime, owing to general hype.
“The hardest change is always bringing people along though. How do you take the 20-year-veteran of marketing, HR, tax; help them learn these technologies and bring them on that journey? Too often that’s triggered by crisis; things coming to a head and forcing change. But we don't want to see crises every day! I think getting people to go on that journey (to adopting new tools and techniques) means helping them recognize that it's not only good for the businesses, it's actually good at the individual level. If you're an accountant, and you don't become the digital accountant of the future, you're really limiting your opportunities.”
“Making change happen is often about recognizing where the individual interests and the business interests are likely very aligned in terms of people going on this journey.”
He adds: “It’s the projects that really impact human life that stand out. An example: we had two hurricanes hit back-to-back in the Caribbean islands. FEMA, the Federal Emergency Response Agency in the US responded to help rebuild. About 140,000 structures had been damaged and you’re not allowed to rebuild until the damage is assessed. The reason being that if more than 50% of the structure is damaged, they want you to rebuild to a higher standard.
"Imagine how long it would take to assess all this, particularly in remote Caribbean islands: it would take over a decade if you were sending in engineers; that's completely untenable. But armed with Alteryx software on a laptop and battery packs because there wasn’t power on these islands, a team from FEMA contractor Atkins went in to respond – and to build models to assess the damage. (The Atkins team used GIS to locate damaged structures, collected information on them and pulled in data from the European Union, NOAA, the National Weather Service, FEMA, and the Army Corps of Engineers, cleaned it, blended it, and created indices in Alteryx.
As Jacobson notes: "They brought in satellite data, data on wind speed, the flood surges, and quickly built models to assess all 140,000 structures, figuring out which ones were clearly well under 50% damaged, which ones are well above 50%... That took years off the recovery effort. And now it's just a standard practice in terms of how we respond to these sorts of things; using data science to assess bed management for hospitals and staffing during the pandemic.”
Driving change through encouraging things like more intelligence use of data does not, in short, always have to be about a struggle to improve the bottom line; it can be about encouraging the upskilling of a single person in HR or accounting, or helping fix tens of thousands of homes after a storm -- and making sure they're built stronger next time. There's a metaphor in there somewhere, for business leaders.
