The Agony and the Ecstasy of Advanced Analytics
Organizations have embraced analytics with gusto. They've cleared away a good bit of the low-hanging fruit, but what's higher up on the tree is going to be much more difficult to get.
- By Steve Swoyer
- March 14, 2016
"What have you done for me lately?"
It's a question that people who develop and implement analytics programs aren't used to hearing, especially from C-level executives. If anything, they're used to being courted, coddled, and indulged.
It's one they'd better start getting used to, however. It's been eight years since industry luminary Tom Davenport fired arguably the first salvo in the advanced-analytics awakening in a now-seminal book Competing on Analytics: The New Science of Winning. Since then, enterprises have taken Davenport's signal insight -- that organizations can, in effect, compete on the basis of analytics excellence -- to heart.
Actually, that's putting it mildly. Organizations have embraced analytics in a way, and with a gusto, that Davenport himself probably couldn't have anticipated. In the process, they've cleared away a good bit of the low-hanging analytical fruit, starting with basic -- but potentially lucrative -- customer segmentation analysis or predictive maintenance. The stuff that's higher up on the tree is going to be a lot more difficult -- and considerably more costly -- to get at, experts say.
This shouldn't -- and won't -- dissuade analytically savvy organizations, they point out. It will, however, change how analytical projects are prioritized, managed, and, most important, funded.
"Over half of [client organizations] are at the beginning of this sort of S-curve of maturity. They're saying 'We need to move quickly. We've been given a budget, but how do we spend it? What's the ... tactical road map? How do we know if our company is ready for this?'" explains Jack Phillips, CEO and cofounder -- with Tom Davenport -- of the International Institute for Analytics (IIA).
"They want to know how many people they need, how they calculate a return on investment, questions like that. Their CFO wants to know these same things. [The CFO is] saying, 'We've changed the culture, we've bought all of these new boxes, [licensed] all of this new software, hired all of these new people. Prove to me that this was a higher return on investment than hiring ten sales people.'
"The highest [analytical] performers we've seen are thinking way out ahead of the CFO's question."
When they say analytics, Phillips and IIA mean analytics -- not basic reporting, not ad hoc query and analysis, not averages, summaries, or basic statistics of or about data. Not business scorecards, performance dashboards, or general key performance indicators (KPI), such as Day Sales Outstanding or Net Margin. All of this stuff is part and parcel of traditional business intelligence (BI).
Analytics describes both a subset of BI and a distinct set of tools, concepts, and methods. It promotes the use of technologies such as statistics, data mining, machine learning, and numerical methods -- i.e., advanced mathematics -- to inform or drive operational decision making.
The vision of competing on analytics that Davenport described -- and which thousands of companies are attempting to make good on -- is one in which analytical insights are pushed down into business processes and delivered on an as-needed, time-critical basis to the frontlines of business. It's a vision in which the data that streams from many different vectors -- from internal OLTP systems and other (stateful) operational applications; from the REST-ful (stateless) cloud; from streaming or event-driven signalers; from the social world of sentiment, intent, and human interactions -- must be synthesized and contextualized into meaningful insights.
It isn't a world for the meek. There's low-hanging fruit, as Phillips and other experts stress, but in order to compete with other companies that are no less determined to win with analytics, an organization must standardize, reconcile, and manage -- in a word systematize -- its analytical programs. "There's one large organization that we work with … and their premise was basically 'I've already picked the lowest hanging fruit off of the tree. In the beginning, I could show 20-times the return [or] $20 dollars of value created for every dollar I invested in analytics, but those days are over. I have to go up higher in the tree now and I'm going after bigger and bigger projects, in terms of driving performance.' What [analytics advocates are] finding is that they're now competing with other functional leaders for mindshare and funding," Phillips explains.
"They have a seat at the table, but they have to put on a whole new game now. The CFO is saying, 'You have to think a lot more clearly about how to articulate the value you bring to the table.' The big question this poses is does analytics have staying power? In order for it to have the same staying power as finance or the supply chain -- departments [the value or validity of which] nobody questions -- [analytics practitioners] are going to have to get a lot more systematic about what they're doing."
Imagination's the Limit
Industry veteran Jill Dyché, vice president of best practices with SAS Institute Inc., identifies three mindsets she says are characteristic of organizations that have invested in analytics.
The first, she says, is when IT leaders believe the analytics investments they've made are sufficient. This "we have what we need" naiveté is particularly dangerous, argues Dyché. IT and business leaders don't even know that there's stuff they don't know. (Nor, for that matter, are they curious about it.)
Less naïve, but no less dangerous, is the second mindset -- the belief that IT can't do more without asking executives for help. In many cases, Dyché argues, this is the product of the specific culture of an organization. IT leaders don't feel they have the organizational authority to ask for help. They're afraid of putting themselves out there only to be told "No."
The third and last mindset is kind of an enlightened version of the first, says Dyché. She likens it to Meno's Paradox, Plato's argument that if you know what you're looking for, finding it is unnecessary, and if you don't know what you're looking for, finding it is impossible. IT and business leaders know there are things they don't know but they don't know what any of it is or what it looks like. They do know they need to find out, though.
This year, Dyché authored The New IT, a book that deals with how IT's identity is changing -- and what that change means for IT's rights, responsibilities, and relationships with other lines of business. The New IT makes the case that IT can play a critical role in enabling organizations to effectively compete on the basis of analytics by building on existing investments and promoting governance and rigor in analytics efforts.
"I would argue that most of the [critical enabling] investment has already been spent on BI and data -- procuring platforms, dashboard and scorecard tools, some light visualizations, specialized skills," she explains, citing data modelers, DBAs, business analysts, data stewards, reporting specialists, data architects, and data warehouse architects, along with mid-level managers, ETL and ELT specialists, and others. In organizations that have made these investments, "advanced analytics investments are merely incremental," she argues.
Dyché returns to Meno's Paradox: as long as an organization knows that it doesn't know something -- as long as it has that driving insight -- it's headed in the right direction.
Beyond this, Dyché argues, there are all-too-human reasons why people themselves tend to resist analytical retrofitting. "Analytics is about giving new news and that's a harder sell because it suggests the possibility of change. And change ... is hard," she says.
"If I can predict that a valuable customer has a high likelihood of committing fraud, what do I do with that? Sure, I could save the company untold money and time by triggering an alert on that customer's account, but I'm not measured on that" -- here she invokes the viral Twitter hashtag #NOTMYJOB -- "so why [should I] stick my neck out if there's only a 72 percent likelihood that he's using an alias to launder money?"
Basic BI: The Gateway Drug to Analytics
Of course, for every organization that's hit an analytics wall, another is just getting started.
The fact is, argues industry veteran Mark Madsen, a research analyst with IT strategy consultancy Third Nature Inc., too many companies are still struggling with basic BI capabilities such as reporting.
"In a general sense, most people aren't using analytics, by which I mean statistics or any other advanced math-y stuff, in interesting ways, but all of the companies that are doing this have definitely been past that point [of gathering up the low-hanging fruit] for a while," he says.
The good news is that even these late-ish adopters are alert to the power and promise of analytics.
"For all the companies that have those advanced [analytics] capabilities, the truth is that so many others are barely even scratching the surface," he continues. "Too many [companies] are still struggling with basic BI, but in a way, they're aware of what they're missing out on because of all of the hype [about analytics]. In some cases, they're coming in with unrealistic expectations, however."
One late-ish adopter is Lynn Community Health, a healthcare provider based in Lynn, Mass. It recently replaced its Excel-based "BI" practice with integrated BI reporting and dashboards, courtesy of a system developed and integrated Quantiphi Inc., a local "decision science" services firm. Quantiphi's implementation is based on Yellowfin, a platform-as-a-service BI offering that can run on- or off-premises. Kate Heffernan, director of decision support with Lynn Community Health, says her organization's immediate priorities were unified reporting and especially dashboard capabilities.
However, Lynn Community Health plans to make use of analytics capabilities -- especially predictive analytics -- as its BI implementation matures, Heffernan stresses.
"We're really hungry to get into more advanced things, [such as] getting into predictive analytics and really starting to be a leader. So often, we're reacting [to things] and we're tired of it," she says.
"In our situation, we're always get stuck ... because the [insurance] payers will come to us and they'll say, 'Your patients are really risky and you're costing us a lot of money. We're only going to pay you this.' [BI reporting] gives us the ability to come back [to the insurers] and say, 'This is what we're doing, these are the patients [we're treating], and this is what we're doing with them.' The predictive part will be the ability to project [likely treatments or medical interventions] based on the information [we have about patients]. This will give us a proactive ability in negotiating with [insurance] payers."