March 10, 2011

The three principal kinds of analytic business benefit

When I tweaked the slide deck for Thursday’s Investigative Analytics webinar — I’ll post an updated version soon — the part that needed the most work was the section on “What business problems do you solve with this stuff anyway?” I’ve posted about that kind of thing at least five times in the past five years, across three different blogs (linked below). But perhaps this time I can really simplify matters, albeit at the cost of being not quite complete.

A large fraction of all analytic efforts ultimately serve one or more of three purposes:

Those areas obviously overlap. Indeed, it can be argued that everything one does in business amounts to “optimization,” and everything in analysis boils down to noticing and understanding anomalies. Still, I am hopeful that this is an instructive categorization, as per the many examples below. 

I go back and forth about whether to include a fourth area, information flow, to cover areas such as accounting, compliance, routine reporting, navigational search, and so on. But ultimately you’re doing those things in the thought that somebody might use the information to make (or at least check) a decision some time. And when the decision is made, it’s likely to relate to one or more of my three initial categories. So this time around, let’s leave that out.

Finally, there are some examples of analytic pattern detection that don’t fit my trichotomy very well, especially in the areas of financial services (algorithmic trading) and research science. Well, as I’ve noted before — no categorization is ever perfect.

Lots of illustrative examples, in (very rough) chronological order

Ever since the breakthroughs of Florence Nightingale, data visualization has served to highlight anomalies, especially ones that indicate problems. (She was focused, as you might imagine, on flaws in medical care that led to excessive numbers of patient deaths.)

Ever since the days of W. Edwards Deming, statistical process control has been used to optimize manufacturing processes by eliminating causes of defects. (That’s what gave Japan the push to leapfrog the US in manufacturing quality.)

Inventory planning has improved for decades, long ago getting past the point of “Oh, we’ll be out of stock soon, better order more.” Depending on industry, inventory planning is a serious exercise in optimization, marketing, or both.

Inventory planning technology is often out of the cross-application analytic mainstream, and the same goes for price-setting. Back in 1984, a United Airlines employee told and artificial intelligence conference that United was making $100 million incremental profit per year via expert-system-aided price-setting. Revenue optimization is, well, a major optimization exercise, with a whiff of marketing.

Hordes of MBAs spend their days in front of Excel or more specialized planning software, optimizing everything they can think of.

Direct-mail marketing offers have long been optimized, first via simplistic A/B testing, then via more sophisticated analytics too. The same now goes for email and the like, and of course also for web page personalization. These are highly important marketing activities, with a strong flavor of optimization. There’s also occasionally a whiff of problem detection, most notably when websites get sudden crashes in their success rates, indicative of technical issues such as slow-loading pages.

An important subcategory of marketing is churn prevention, which is often an exercise in discovering particularly bad problems in customer satisfaction. Social media analytics also can introduce a strong element of problem diagnosis — sources of spreading dissatisfaction — into marketing.

Increasingly many investment decisions are made via algorithmic trading. Sometimes such algorithms directly look for early evidence of problems. Even when they don’t, they could vaguely be described as the detection of problems (pricing “errors”) or as portfolio optimization. But frankly, this is one area that isn’t really a good match for my trichotomy. On the other hand, the related and burgeoning area of risk management fits pretty well into the “problem diagnosis” bucket.

Anti-terrorism is all about detecting some rather serious problems. So, less dramatically, is anti-fraud.

Earlier tries at this kind of categorization

I’ve taken whacks at this kind of breakdown a few times before. Back in 2006 I rattled off a long list of early-warning uses for text analytics. The same year I discussed application areas for data mining and came up with a list much like the one in this post — lots of early-warning or other problem detection, lots of marketing, plus the scientific and algorithmic trading examples also noted above. Also the same busy year, I segmented business intelligence into four categories:

Flashing forward to 2009, I unearthed a list of specific marketing uses for analytics, originally compiled by Mike Ferguson. That same post starts with a Teradata-supplied list of cases in which you’d want the benefits of your analytics to be delivered near-real-time. And finally, a few months ago, I opined that text analytics application areas typically fall into one or more of three broad, often overlapping domains:

Comments

4 Responses to “The three principal kinds of analytic business benefit”

  1. Notes for my March 10 Investigative Analytics webinar | DBMS 2 : DataBase Management System Services on March 10th, 2011 3:24 am

    […] Here’s a more current version. A number of the changes arose when I thought more about how to categorize analytic business benefits; hence that blog post a few minutes ago with more detail on the same […]

  2. John M. Wildenthal on March 10th, 2011 10:21 am

    It isn’t anomalies that are important. It is consistencies. If we aren’t starting with a consistency, we cannot extrapolate to what is an anomaly. Actionable analytics is about relationships – consistent and persistent relationships. A cluster anomaly can become important because it exhibits a consistently different relationship than the rest of the data. Individual anomalies may not be as interesting as they may just be noise or truly rare events. When a cluster develops, that is when it gets interesting.

  3. Curt Monash on March 10th, 2011 12:31 pm

    John,

    That’s a fair point. Thanks!

    My slide deck version of the story makes “patterns” more central to the story than what I posted here, which is probably more in line with what you’re suggesting.

    When one focuses on patterns, one focuses on discovering new patterns one hadn’t seen before. I was thinking of those as “anomalies”, but perhaps being a little careless in doing so.

  4. The three principal kinds of analytic business benefit | DBMS 2 : DataBase Management System Services « databasing on April 23rd, 2011 8:53 pm

    […] The three principal kinds of analytic business benefit | DBMS 2 : DataBase Management System Service… Posted by databasing Filed in Uncategorized Leave a Comment » LikeBe the first to like this post. […]

Leave a Reply




Feed: DBMS (database management system), DW (data warehousing), BI (business intelligence), and analytics technology Subscribe to the Monash Research feed via RSS or email:

Login

Search our blogs and white papers

Monash Research blogs

User consulting

Building a short list? Refining your strategic plan? We can help.

Vendor advisory

We tell vendors what's happening -- and, more important, what they should do about it.

Monash Research highlights

Learn about white papers, webcasts, and blog highlights, by RSS or email.