Tom Davenport

Private Equity and the Ownership Decision

2009-06-29T15:30:29Z

A few months back I wrote about micro-decisions and how they can add up to macro money. But as a student of decision-making I would be remiss if I didn't point out that one big decision — who will own an organization — has a massive effect on performance, culture, and organizational effectiveness.

I'm prompted to post on this topic because of two contrasting experiences of late. One is that I've been doing some research on private equity and what a bloody mess it has made of many of the companies it has taken over. The other is a recent visit to SAS, the privately-held company that dominates the market for analytical software.

Private equity is in shambles now, and deservedly so. Deals are down 96% from their peak. The "industry" has run several pretty good companies into the ground, including Mervyn's, Linens 'n Things, Steve & Barry's, Station Casinos, and 66 of 105 bankruptcies in 2009. BusinessWeek had a great story on the demise of Mervyn's, which can be traced to a combination of PE greed, overly creative asset management, and a somewhat outdated business model for the underlying business. A typical PE anecdote: while Mervyn's owners took $137 million in "distributions" over the first two years of the deal, water coolers for employees were taken away.

Sure, some of the retailers involved in PE deals would probably be having hard times anyway. But previously healthy companies can be brought down by PE too. Take Harrah's, for example, which has been the poster child for effective use of analytics in its business. The gaming giant went private in a 2006 PE deal with Apollo and TPG. They saddled it with nearly $25 billion in debt. Even though the company is still relatively healthy in operational terms, the debt load is crushing. What a shame to see a previously well-managed company stumbling to such a degree.

Contrast all this with the state of SAS, a company I've done a lot of work with over the past several years and have grown to respect greatly. SAS is privately held by its two founders, Jim Goodnight and John Sall. Goodnight runs the company with a "get rich slow" approach, and indeed that's what has happened. The company had revenues of over $2.25 billion in 2008, and seems to be doing well in the recession. Goodnight invests about a quarter of revenues in R&D every year, and SAS is well-known for its fine facilities and services for employees (and consultants — the Umstead Hotel on the company grounds is fantastic). Goodnight also invests heavily in education for the Raleigh area, and has recently built a solar energy farm on the SAS campus.

Which would you rather have; owners who load your company with debt and take away your water, or owners who give you gyms, on-site subsidized childcare, and a piano player at lunch? Granted, not all private owner/operators will act like Jim Goodnight, but it's hard to imagine private equity owners investing for the long haul and to keep employees productive and happy. Think about this decision before you sell your company to a PE firm — or work for one that's already been sold.

Why 1.5 Is Greater Than 2.0

2009-06-15T13:42:29Z

The proponents of 2.0 thinking on user-generated content, be they fans of Web 2.0, Enterprise 2.0, Health 2.0, or Rhubarb 2.0, would have us believe that their highly participative approach is the only one that works. And indeed, there is an appeal in democratizing content creation and management. However, in almost every case there is also value in professional involvement.

Take health care, for example. A couple of months ago there was a conference in Boston (I was unable to attend) on "Health 2.0 meets Ix." For those unfamiliar with this debate, Health 2.0 fans advocate patients taking control of their own health care and sharing information across patient communities, rather than turning it over to professionals. "Ix" refers to "information therapy," which is shorthand for the scientific/medical establishment's 1.0 use of research, clinical trials, and licensed practitioners to fight disease. If you were sick, would you rely on Health 2.0 or 1.0?

In all likelihood, you'd go for both. You'd listen to what the medical establishment prescribes for your ailment, but you'd probably also check out blogs, wikis, and other patient-generated content and communities. As health care analytics expert Blake Zenger notes in his blog, Health 2.0 versus Ix is a false dichotomy. We can embrace the virtues of democratized health content without throwing away the benefits of professionalism and science. Depending on the circumstances, your personal equation may be 1.3 or 1.8, but you're going to want both medical science and the comments of those who have opined on their own health situations.

The same is true in many other content settings. In journalism, while it's useful to experience the passion of bloggers at the Huffington Post and the Daily Beast, they don't replace the reporters and editors at the New York Times. Even Wikipedia, often held out as the epitome of 2.0 content, has increasingly employed (mostly unpaid) editors to monitor, verify, and sometimes even create content.

Inside enterprises, the same 1.5 mix is often desirable. For example, I once heard Steve Schmidt, the CIO of Vertex Pharmaceuticals, describe his company's use of wikis for capturing the results of research. He said that the most successful ones are "curated" — facilitated and edited by humans whose job it is to do so. Before blogs and wikis came along, the same was true of discussion databases. The best ones typically had facilitators and online community organizers.

Of course, it's more romantic and revolutionary to assert that only the masses can generate useful content. It's appealing that the hoi polloi can replace experts, editors, and experienced professionals. It just doesn't happen to be true. The key word is "augment," not "replace." 1.5 is greater than either 1.0 or 2.0.

The Return of the Non-Virtual Organization

2009-05-29T17:56:20Z

I can't tell you how many companies I have worked with that have encouraged or tolerated a large degree of geographic dispersal among employees and management teams. "We're virtual, and proud of it," one told me. "It doesn't matter where you live anymore," many employees of virtualized companies have argued. "We travel all the time anyway," has been another frequent mantra.

But I recently encountered a company that is moving the other way. Eclipsys makes software for healthcare providers. The company's headquarters is in Atlanta. Last week, it changed CEOs. The previous CEO, Andrew Eckert, lived in Silicon Valley. By all accounts, he did a good job in the role, and the company has been doing well. However, the board of directors felt that the company couldn't be managed successfully from afar, and held discussions with Eckert about moving to Atlanta. He was committed for both family and career reasons to stay in California, however, and declined to move. The board decided to change leaders, and Philip Pead, who had previously headed and sold a healthcare software company in Atlanta, got the nod as the new CEO. Pead had moved to Miami, but is returning to Atlanta to run the company.

Pead said this week in an address to customers, "You can't deny how effective it is to be able to sit down and have lunch with another leader and resolve an issue quickly." My sense is that he's right and we all know it. However, many companies seem not to want to acknowledge it.

Of course, virtually every large company has some degree of geographical dispersion. Several Eclipsys managers told me that if they insisted on everyone living in Atlanta, they'd lose a lot of great people. And because the company grew partly through acquisition over the last several years, there are several natural centers where employees are based. I don't get the feeling that there will be massive consolidation at the company, but it seems likely that the top management team will eventually be in the City Too Busy to Hate, as they call Atlanta (I grew up in the nearby city of Birmingham, AL, which was sadly not too busy to hate).

Senior managers, in particular, are a group that benefits from high-bandwidth interpersonal contact. Henry Mintzberg and other researchers have shown that their jobs typically consist of a variety of short, and frequently unplanned, interactions. It's much easier to accomplish these when you are all in the same vicinity.

Senior managers are not the only group I've heard about that are coming back into co-located offices, however. The big push that IT firms such as Sun, IBM, and AT&T made a few years back toward virtual offices for everybody seems to have been dialed back. Many companies allow some work at home, but far fewer seem to support it for five days a week, 50 weeks a year. It seems that "Out of sight, out of mind" has prevailed over "Absence makes the heart grow fonder."

What do you think about this? Does Eclipsys represent the beginning of a trend, or an outlier? Do you see your organization de-virtualizing in the future, or are you still committed to living the virtual life?

The Rise of the Chief Performance Officer

2009-05-11T14:39:53Z

You may recall the brief run of publicity for a new role in Washington called the Chief Performance Officer. President Obama planned to nominate Nancy Killefer, a McKinsey consultant, to this position, but some tax issues led her to withdraw her nomination. A few weeks ago Obama nominated Jeffrey Zients, another consultant and Washington business executive, for the role. I don't know Zients, but I think the Chief Performance Officer role has a lot of potential, and it's a new wrinkle for something like this to appear first in the federal government.

A little background behind the appeal of this idea: for several years at Babson we've run a couple of sponsored research programs — one involving knowledge management and one based on process management. I've had a long-term interest in both topics, but each is limited as an approach to business performance improvement. We had a meeting last week of the knowledge management group. At that meeting, we advocated for merging knowledge management with some other function — most likely the human resources/organizational learning/talent management constellation. We felt that knowledge management groups don't often have the critical mass to stand alone, and knowledge and learning are very similar concepts anyway.

Some of the participants pointed out that if you're going to be merging things, you might as well go a bit further. They noted, for example, that if you want to align knowledge and learning with work, you need to know something about business processes and how to improve them. And if you're going to align processes with the content needed to perform them effectively, you need to know something about the technology that would deliver the content in accordance with job tasks.

What this begins to suggest is that the era of siloed business improvement activities will give way to applying a variety of interventions to improve work. In other words, an organization run by a Chief Performance Officer might be called for. Such a group might have a very broad toolkit — from knowledge to learning to process interventions, and perhaps IT as well. Particularly with knowledge work, these interventions tend to come in linked forms.

While I don't think there are a lot of Chief Performance Officers yet (Wikipedia mentions one at Yahoo, but I gather that role is really focused on website performance), I do think this broader conception of business improvement will catch on. At Johnson & Johnson, for example, the former Process Excellence organization has migrated into a Business Improvement Services organization, which works with J&J business units on process, knowledge, information, and decision-making topics.

The danger, of course, is that a broad business improvement organization would have such breadth that it would lose focus, or that no individual business improver could master the broad array of tools offered. It will be interesting to see how organizations resolve this tradeoff of capability and focus.

How about your firm? Are you seeing a broader business improvement focus? Do you want to see a Chief Performance Officer in your organization? Do you want to be one?

Will Monster.com Go the Way of Newspapers?

2009-04-29T18:08:43Z

I was fascinated by a recent article in the Boston Globe about the factors behind its own impending demise. The article revealed that in 1995 Jeff Taylor, the founder of Monster.com, approached the Globe with an offer to collaborate on online job advertising, and for the Globe to invest in Monster. Needless to say, the Globe's owners (then a different Taylor family) turned Jeff Taylor down. Also needless to say, the Globe has suffered greatly from the loss of classified job advertising that was at one point its annual $100 million cash cow. Most of those fish have now moved to online fishing grounds, including Monster.

Granted, it was 1995 and the Internet was just emerging as a disruptive force. However, I suspect that Monster's ability to search across jobs and match jobs to applicants was apparent even then. The story suggests that it's extremely difficult to embrace threats to an existing business and to anticipate the direction of customers and business models. It even seems difficult to follow the fish once their direction is clearly established.

It's easy to criticize the Globe's conservative owners, and to attribute the current position of newspapers on the precipice of disaster to shortsightedness. However, some of Monster's own decisions also illustrate the difficulties of following the fish.

According to my sources, managers within Monster suggested that the job board site was itself under threat by the next generation of job-finding tools. So-called job aggregator sites such as Indeed, SimplyHired, and Jobster allow the viewing of available jobs across a variety of job boards, and also list their own jobs. While none — alone — is as popular as Monster, together the aggregators get more traffic. Indeed.com is growing faster in unique visitors than Monster.

It makes sense, of course: why not find out about as many jobs as possible? Of course, the fastest-growing online job sites are the free ones: Google Base and Craigslist. It may be lowest-common denominator functionality, but that never seems to stop a business model from advancing. (Neal Bruce, a former Monster employee who is now with First Advantage, talks about this change in seeker behavior in a recent interview.)

Monster made the same mistake as the Globe: they believed their business model was inviolate. They didn't anticipate, and didn't recognize, that the fish were swimming in a new direction. This illustrates not that either company's managers were stupid, but that it's very difficult to act on threats to the existing business model.

How do organizations fight this tendency? One way is to create an influential group of employees (or managers or consultants) whose job it is to advocate for potential threats and business model alternatives. Their job is to persuade senior executives that their current business model will soon be defunct. A variation on this model, as Clay Christensen suggests, is to let the group actually start or invest in the new business model. These approaches will not always succeed, of course. But they might increase the chances that an organization can successfully follow the fish.

A Safe(r) Financial World Order

2009-04-15T18:32:07Z

Okay, it's fun to tweak Twitter and the Twitterati, as I did in my last post, and it's great that such remarks bring out so many commenters. This week, however, I'm writing about a topic that is much more important to us all — even if you think Twitter is the greatest thing since CB radio. The topic for today is how financial services organizations will handle complex analytical decisions in the future.

Why is this topic more important than trillions of tweets? You may recall that it has recently almost destroyed our financial system and our economy. How our financial institutions made decisions about their operations and customers is at the heart of the financial crisis we're in. Managers of these institutions entrusted decisions to analytical models they didn't understand. Regulators allowed banks, rating agencies, and nonbank financial firms to make millions of bad decisions. Financial innovation led to financial irresponsibility.

Over the next several months and years, regulators and management teams will be putting in place systems, processes, and policies to try to prevent rampant bad decisions. What should be the attributes of these controls? I have a few ideas:

Transparency. It's clear above all that future models will need to be more transparent. Managers, regulators, and in some cases customers will need to know more about how investment and risk decisions are made. If it's not clear how a mortgage underwriting model works, for example, then a regulator shouldn't allow it and a manager shouldn't let it be used. I'm guessing that greater transparency will be a key focus of Basel III regulations when they're developed.

Every manager a quant. Since managers are going to have to understand financial models before approving them, they're going to have to learn about quantitative analysis and how models work. Robert Shiller, the Yale economist who is one of the most helpful in his profession at explaining the crisis, comments on this issue in a recent McKinsey Quarterly interview:

"You have to be a quantitative person if you're managing a company. The quantitative details really matter. And so you're going to have to be looking at models, but I think that part of what brought us into this crisis...[is] a kind of a blind acceptance of others' authority that, in terms of variance [value?] at risk or econometric models, or in terms of the counterparties we were dealing with, we tended to think of as bigger than life. And so, people were taking too much for granted.'

Know the flow. A lot of financial models and decisions are baked into computerized workflows. If you want to know their influence within the organization, you need to know where the workflow goes and what the branching conditions are. For example, is a credit model used only for initial loan originations, or for refinancings too?

No more missing data. Missing data, such as documentation of a consumer's income, started as an exception but became a regular practice for many lenders. We know what resulted from the practice. It's not likely that most careful bankers will allow such missing data in the future.

Risk matched with opportunity. For too long in financial systems, risk and opportunity have been separated from each other. Each loan or deal involves both. It's crazy to think about either individually — e.g., the aggregated "value at risk" models — without knowing about the other side. Lenders, investors, and rating agencies kidded themselves when they argued that high return could be achieved without high risk.

Assumptions surfaced. Every quantitative model involves assumptions — about the context in which it applies, the data from which it is drawn, and the underlying basis for the behaviors it predicts. Some typical assumptions might be, for example, "This mortgage underwriting model really only fits markets where housing prices are continuously rising," or "This credit card pricing model assumes chargeoff rates similar to what the bank experienced between 2000 and 2005." You can't really understand whether to use a model unless you know the assumptions behind it. Any model developer should be forced to list the assumptions behind each model, and the model should never travel without its accompanying assumptions. Of course, people can still make stupid decisions once they know and acknowledge the assumptions, but it becomes a bit harder.

Of course, the general subtext of all of these trends is a more conservative set of decision approaches. The key is to employ these more conservative approaches while not stifling innovation in financial products and processes.

Is Twitter for Serious Marketers?

2009-04-09T15:38:56Z

A few months ago I was speaking at a marketing conference, and after I spoke on marketing analytics, there was a panel on social media. Larry Weber, who started and then sold a very successful PR firm (and who is on Babson's Board of Trustees), was asked whether there was a role for analytics in social media.

"Frankly, I'm tired of analytics," he said. "I got into social media in part to get away from analytics." Well, honesty is good, but I didn't see then — and don't now — how you can do serious marketing through any medium without metrics and analysis. Twitter and other social media may be fun, but are they really serious marketing tools?

I thought of this again recently while grading some of my MBA students' papers about an IT strategy for Welch's, the grape juice people. A couple of the student groups suggested that Welch's should embark upon a Twitter initiative. Okay, they get a point or two for being au courant. And to the students' credit, most suggested that it was a low-risk, low-return marketing approach. Still, I couldn't imagine which customers would decide to follow Welch's tweets about its grape juice and other associated products. The busy moms who form Welch's core customers? I don't think so.

Do serious marketers spend a lot of time and energy on Twitter campaigns? I doubt it. Sure, go ahead and play around with it — it doesn't cost much. But I defy you to do serious brand management in 140-character messages. I defy you to prove that Twitter users are your typical customer — unless you sell bubble tea or something similar — or that their tweets are a true reflection of their relationship with your company.

Let's face it — Twitter is a fad. It has all the attributes of a fad, including the one that people like me don't get its appeal. It has risen quickly and it will fall quickly. It's this year's Second Life — which, you may have noticed, nobody is talking much about anymore. One Daily Telegraph article that did talk about it noted, "While the site is still beloved by geeks and the socially awkward, Deloitte's director of technology research, Paul Lee, says it has been "virtually abandoned" by "normal" people and businesses." Ouch!

I had a conversation with an influential business editor the other day that confirmed some of my predilections about Twitter. He said he was "unfollowing" (defollowing?) those who tweet a lot — "It's just become a burden to read them," he said. I, who issue nary a tweet, am clearly sitting in the catbird seat. You have to wonder about a technology when those who use it aggressively are shunned.

I'm not as negative about the business and marketing potential of some other social media. For example, because Facebook and MySpace offer the promise of monetizing social networks — though they haven't done so yet, to my mind — they are not to be easily dismissed. And wikis clearly have some value, or Wikipedia wouldn't be so useful. Yet I haven't seen too many wiki success stories within firms, and the ones that do have value don't involve marketing. One smart knowledge manager, Sukumar Rajagopal at Cognizant, told me that he thought successful wikis within companies required that participants in them have strong network ties, and that's not always easy to orchestrate. Another pharma executive who had experimented with them suggested that they require substantial human curation (facilitation and editing) to be successful — which, come to think of it, Wikipedia does too.

One conclusion I've come to is that we should unbundle the concept of "social media," because some of its components are much more useful than others in a business and marketing context. Facebook? I suspect it faces prosperity, over time. Second Life? On life support. Twitter? In the long run, not worth a tweet.

What do you think? I'd love to hear your thoughts, but please restrict them to more than 140 characters.

For an opposing view, see John Sviokla and Chris Curran's post, "Twitter: A Marketer's Duct Tape"

Is Forced Time Off Fair?

2009-03-16T19:30:41Z

One of the common approaches to dealing with this recession is for companies to ask — well, tell — employees to take time off without pay, a day every week or two. This 10 or 20% haircut is supposed to indicate that "we're all in this together," and that it's better for everyone to suffer a little than to lay some people off. While I have some sympathies with this philosophy, I'm not sure it's either fair or wise.

On the issue of fairness, if such a policy had been instituted in 1969, it might have been very fair. But in 2009 there is much less of a relationship between hours on the clock and work actually done, at least for knowledge workers. How many of you reading this post actually work only 40 hours a week? How many of you only work on official workdays? Today, most people have a continuous mixture of work and non-work activities, and it will be difficult for any knowledge worker to stop working for a day every week or fortnight. I might suggest that this is exactly what the employer wants, but that would be a cynical remark.

There is also the issue of whether the forced haircut is wise. I have problems with its wisdom in two respects. One involves the fundamental principle that all employees are equally valuable. It's nice to pretend that they are, but we all know they're not. Giving all employees a haircut may lead the most valuable ones to look elsewhere. There was a column in a recent Boston Globe about treating all employees (at Boston's Beth Israel hospital) alike with regard to cuts. It's heartwarming, but if it leads to an across-the-board haircut, might some of the best employees leave for wealthier hospitals across town?

The other potential problem is that employees, given an involuntary time chop, may look elsewhere to fill the void. They'll freelance, e-lance, or moonlight to replace the lost income. This could lead to a variety of negative scenarios for the employer/barber who originally chopped their time. The employee might find the freelance employer more desirable, and jump ship altogether for full-time employment there. Or he might end up doing a bit of his freelance work while ostensibly on the clock for the 80% or 90% employer.

I'm not saying that 10 or 20% haircuts for everyone are necessarily a bad idea. I do think, however, that they are hardly a no-brainer either. The inclination to share the pain is admirable, but it could open the door to a host of problems.

Microdecisions for Macro Impact

2009-03-04T20:44:59Z

When companies or pundits talk about decisions, it's easy to get focused on the major ones made by senior executives. It's true that big decisions around restructurings, mergers or acquisitions, and strategy can create or destroy a lot of value. However, those decisions don't typically happen very often, and they are not always amenable to improvement. The senior managers who make them, for example, may not want you mucking around in their decision processes.

What many companies don't realize is that microdecisions — small decisions made many times by many workers at the customer interface — can have a major impact on the business. How they are made can be the difference between sloppy and effective execution, and between profit and loss.

If you can identify a few key microdecisions that can be addressed and improved, you can often dramatically improve performance.

What are some examples of important microdecisions? It depends on the industry, of course--these decisions are found deep within key operational processes. If you're an insurance company, for example, day-to-day decisions about how claims are settled can make or break your performance. If you're in banking, we now know that loan decisions are critical not only to profits but to our entire economy. Increasingly, banks are having to make important microdecisions about delinquencies and collections. If you're in health care, key microdecisions involve not only the clinical treatments administered to patients, but the daily decisions about how long patients stay in hospitals and how their bills are paid. For a drug company, I once did a consulting project that found that the length of time a physician prescribed a drug was one of the most important microdecisions affecting the company's profitability. And in almost all industries, the price that businesses charge for their services is one of the most important microdecisions of all.

Let's say you've identified a microdecision or two that has economic leverage. What can you do to improve it? There are many possible interventions, and it's important not just to always use the same one. One approach is to automate it entirely. This is the focus of James Taylor and Neil Raden's book Smart Enough Systems, and of Taylor's blog on enterprise decision management. . If the decision is structured enough, that may be a good idea.

If you're pretty sure that you will still need humans in the process, there are again several possible approaches. One is to establish key metrics that are consistent with the goals of the process, and begin measuring the performance of the human decision-makers on them. How do their decisions turn out? How do customers feel about the service they received? Until we start to measure micro-decision makers at the individual level, we aren't likely to improve their performance much. This is the approach of a company called Athenium, which applies it regularly to insurance, health care, and legal processes. (Having breakfast this week with Andy Snider, the company's Managing Director, reminded me of the importance of these micro-decisions.)

One of the simplest approaches to improving micro-decisions is to give those who make them a checklist to make sure they didn't forget any key steps or issues. I have previously blogged about this method, as described by doctor-writer Jerome Groopman. Atul Gawande, a surgeon in Boston, has also written persuasively about checklists in The New Yorker, and he's co-author of a recent New England Journal of Medicine article concluding that a 19-item checklist reduced deaths during surgeries by almost half. Pilots, of course, have used checklists for decades to decide whether their planes are ready to take off.

What microdecisions has your organization addressed and improved? And how have you improved them? Most importantly, has it made a difference in your organization's performance?

What HR Analysts Can Learn From Basketball

2009-02-18T14:25:34Z

Michael Lewis has done it again. In an article in last Sunday's NY Times, the Moneyball author applied his considerable writing skill to describing the penetration of analytics into basketball. As in his baseball bestseller, where he focused on the Oakland A's, the article focuses on particular players (Shane Battier of the Houston Rockets) and managers (Daryl Morey of the Rockets) to show how a scientific approach is changing the game.

Lewis didn't go into the details of basketball analytics (for that see, for example, 82games.com), but he does illustrate how a few leading individuals can change how a major sport is viewed and pursued.

If you don't care about professional basketball, is there any reason why you should care about the use of analytics in it? Yes -- because analytics in sports catches on and spreads in a similar fashion to business. Also, basketball is more like business than most sports, and the analytical approaches there have more potential relevance to businesspeople.

How do analytics spread in sports? It usually starts with a few individuals who have seen their application in other domains (Daryl Morey of the Rockets, for example, was a fan of Bill James, the baseball stats Geek of Geeks), and figures they will work in a new context. Some like-minded rich people bankroll the experimentation (in the Rockets' case, owner Leslie Alexander), and the team starts to perform pretty well (Houston had a 22-game winning streak last year despite injuries to key players). New metrics get developed--both by teams and amateurs outside them. Then other teams catch on. The last time I checked about a year ago, roughly half of NBA teams had statisticians on staff.

Except for the development of new metrics by amateurs, the same adoption approaches apply in business. Capital One, Harrah's, and Progressive Insurance play the role of the Houston Rockets and the Oakland A's. Capital One's Rich Fairbank and Harrah's Gary Loveman are the individuals with the vision; Signet Bank and Harrah's Phil Satre bankrolled their respective ideas. And in many cases, the ideas were already being applied in other industries. Marriott, for example, adopted analytical revenue management in part by observing the same approach in the airline industry. And as with sports, in these industries almost every major player has adopted analytical approaches by now.

Basketball is more like business than many other sports because of the importance of team play and interpersonal dynamics. Houston's Battier doesn't score or rebound a lot, but his team clearly plays better when he's on the court. A group of statistical metrics called "plus/minus" compare team performance when a player is in the game versus when he's out; Battier excels on that metric.

In business, we've all known managers whose units or companies perform better when they're in charge. Unlike professional basketball, however, most companies haven't yet begun to evaluate managers or employees systematically based on their individual and team contributions. No plus/minus statistics have been developed for a business context. The emerging field of human resource analytics has a lot to learn from the Houston Rockets.

* * *

On March 10th, Tom Davenport will be hosting a webinar on winning with analytics.

Why Are Math Jobs So Much Fun?

2009-01-30T01:51:49Z

I've been a little slow to blog lately because I'm trying to finish a book on how organizations can become more analytical. But recently I was alerted (by Al Parisian of Montana State Fund) to a ranking of the best jobs by the online job site CareerCast.com. I (and my co-authors Jeanne Harris and Bob Morison) have been writing about the value of hiring analytical people for the companies that hire them. But it turns out that the jobs are also great for those who have them.

The top three jobs in the survey--mathematician, actuary, and statistician--are all highly quantitative. Several others in the top ten--biologist, software engineer, computer systems analyst, and sociologist (yay--that's my field)--are also often mathematical.

Why did the math-oriented jobs rank so highly? Any analytical reader knows to look behind the rankings for the methodology. In this case, 200 jobs were ranked on their relative income, work environment, future employment prospects, physical requirements, and stress. Numbers-focused work tends to be highly paid, is done in safe work environments, will be highly sought-after in future labor markets, involves no heavy lifting, and isn't that stressful--at least, if you like math. So it's no wonder that the jobs were highly ranked.

One wonders, then, why American students haven't gotten the message about how good these jobs are, and why they haven't started hitting the books in math courses. The TIMSS report on a 2007 study of global student math and science performance suggested that while the US has made some strides, it's still well behind a number of Asian countries. There is fairly universal agreement among technology executives and knowledge work gurus that this problem will have to be solved if the US economy is to prosper.

One answer may be that math students aren't that good because their teachers aren't that good. One recent study suggested that many math teachers aren't certified to teach their specialties. And other recent studies--as Bill Gates argued in his first foundation letter -- suggest that good teachers make an enormous difference in how successful students are. Maybe we just need better math teachers.

If you're already a math teacher, you could start motivating your students by telling them what great opportunities there are to use math in the adult job world. And it might be useful if parents employed some of the same arguments. I think most good jobs in the future will involve some sort of math or analytical component, so we'd better get our children ready. What do you think we could do to increase the mathematical orientation of today's students?

* * *

On March 10, Tom Davenport will participate in a webinar on how to win with analytics.

The Year Ahead: Make Better Decisions

2009-01-05T13:00:00Z

2008 and several years preceding it will not go down in history as the best years for good decisions. While we did make a groundbreaking (and, so far it seems, pretty good) decision to elect a different kind of man as U.S. president, we have made a lot of bad ones, both individually and organizationally. We invested poorly, we committed resources in the wrong places, some of our friends turned out to be not so friendly, and we didn't prepare for tough times.

So let's make 2009 the Year of Better Decisions. In what better way could we improve our personal lives and our organizational effectiveness? In the rest of this post, I'll describe what it means to focus on making better decisions.

One key step for individuals or organizations is to make a list of key decisions--key by whatever criteria the person or organization cares about. It could be the "five key decisions for my financial future" or the "top ten decisions required to execute our strategy" or "the top 20 decisions that have to go well for us to meet our financial goals." Without some inventory, all decisions will be treated as equal--which probably means that decisions won't be addressed at all.

In addition to an inventory of decisions, it's important to classify each major decision by its type. Is it financial, personal, strategic, or tactical? How frequently does it recur? How structured is it? What information is available to support it? Such a classification would begin to allow an organization or an individual to understand what interventions might make the decision more effective, and would establish a common language for discussing a decision.

Decision-oriented firms and individuals would probably also have some way to track decisions and their outcomes. This can get a little political, so few organizations do it yet. But it's probably coming. Eventually this will probably be possible in software, but for 2009 feel free to use paper and pencil.

Individuals who want to make better decisions probably have to rely on themselves, but a decision-focused organization would probably have a group of decision "engineers"--or coaches or consultants--whose jobs involved improving decisions. At GE Money, for example, a group of about 400 analysts reside within a "Decision Management" organization. Instead of just supplying analytics and correct answers, the goal of the analysts is to work with executives to improve decision processes.

There is one other key attribute that will need to happen if decisions are going to be improved. It's a key behavior that is fundamental to any decision intervention. I call it "meta-decision analysis." It simply means that before making a decision, a person or organization should ask, "How should we make this decision?" At Air Products and Chemicals, for example, key decisions are expected to follow a five-step process.

Step 1 is to define the decision to be made.

Steps 2 and 3 are what might be called meta-decision analysis; they are to "determine method" and "establish governance." The method to be adopted involves the level of participation; options include:

Unilateral
Consultative
Majority
Consensus

The governance approach in the decision analysis follows the well-established RACI approach in project management: who is expected to be responsible, accountable, consulted, or informed?

Step 4 in Air Products' approach is to make the decision; Step 5 is to communicate and implement it.

Why is such a meta-decision approach important? It's because there are many different ways to make decisions today, but they all require stepping back and thinking about the decision process and the best ways to accomplish it. Air Products' approach is far better than what most organizations do, but there are many more possibilities and options to choose from. Just the meta-decision about how participative should the decision process be, for example, can require a great deal of analysis. Victor Vroom, a Yale School of Management professor, has devoted much of his academic career to the study of that issue. He's developed a meta-decision framework to help managers decide how participative their decisions should be. The framework addresses issues such as:

How significant is the decision?
How much does the decision-maker know about the decision?
Are stakeholders likely to commit to the decision if they don't act on it?
How important is speed in the decision process?

This participation issue is one that works at the individual level too; for example, it's probably a very smart idea to consult your spouse on investment decisions!

Vroom even developed a software program that takes managers through a set of questions and then recommends a level of participation. This would be a useful tool, but the problem is that participation is only one variable to be considered in meta-decision-making. How about the information to be used in the decision? Should the decision itself be automated, or should it rely on human brainpower alone? Should a prediction or opinion market be employed? Should some sort of devil's advocate be employed? That technique has been shown to be a useful addition to many decision processes. What other roles should be adopted, and who should play them?

Eventually there will be software tools that help us manage and improve our decision processes. Don't look for them in 2009, but it probably won't be long thereafter. For the year ahead, however, don't be held back by technology--do what you can with manual systems.

Can Millennials Really Change the Workplace?

2008-12-15T15:50:04Z

I wrote a post a few weeks back about whether the current economic climate might lead to a somewhat reduced focus on twittering and friending a la Web 2.0. I thought my comments were fairly mild, but apparently not in the view of many commenters. I haven't been accused so vehemently of not understanding the younger generation since I insisted that my teenaged sons go on vacation with the family.

Now I am fairly secure in my membership in Generation Jones--I still have lots of pleated pants, for example--and I find the accusations that "I just don't get it" amusing. But the comments did make me think about the fate of the millennials as they move into the workforce. Will they bend to the whims of the workplace, or will the workplace bend to suit them?

Of course, we don't know for sure which sort of bending will eventually win out. But there are some fun clues from two different settings. One is the award-winning AMC hit series "Mad Men," in which the work and lives of 1960's advertisers are fictionalized. I'm watching the second season now on my DVR; the first season is available on DVD. The ad agency chronicled in the series, Sterling Cooper, has decided that it needs to "think young," and has hired some youthful employees to represent their generation to clients. One of them, Smitty, has lines like, "Our generation doesn't want to be told what to do or how to act. We just want to BE," and "Stop telling my generation what to do, man. We want to find things for ourselves, dig? We want to feel." His actual achievements are somewhat less radical; he comes up with a new ditty to sell coffee, for example. Of course, we don't know whether Smitty eventually sells out, but the fact that he works in an advertising agency suggests that the chances of his revolutionizing the workplace are slim.

Wall Street Journal columnist Ron Alsop's recent article in the WSJ , "The Trophy Kids Go to Work," includes quotes from millennials that are reminiscent of Smitty's. The article, derived from Alsop's new book, describes some of the work-related attitudes of the generation that got trophies for just showing up at soccer.

Now I dislike these generational generalizations, but Alsop has come up with some interesting observations. I loved this quote, for example, from Olivia, a teenage blogger: "They are finding that they have to adjust work around our lives instead of us adjusting our lives around work...What other option do they have? We are hard working and utilize tools to get the job done. But we don't want to work more than 40 hours a week, and we want to wear clothes that are comfortable. We want to be able to spice up the dull workday by listening to our iPods. If corporate America doesn't like that, too bad."

It's easy to lampoon remarks like these, and I doubt that Olivia will be completely successful in her battle with corporate America. I see a lot of organizations that have banned Facebook and Twitter access at work, for example, and they still seem to be able to recruit young workers. The CIA even bans mobile phones in the workplace (supposedly for security reasons), yet young college graduates are flocking to the place.

But I hope Olivia and her generation are at least somewhat successful in bringing about a better working environment. The workday is dull, most people do work too many hours, and suits and ties really aren't very comfortable.

The 60's generation largely restricted its social change to nights and weekends. I hope the millennials change the workday from 9 to 5.

10 Principles of the New Business Intelligence

2008-12-01T15:36:24Z

Business intelligence--and its predecessor concepts decision support, executive information systems, and so forth--have been circulating for several decades in business. However, I don't think it's ever fully worked. What we've done is to throw data (often in the form of difficult-to-navigate data warehouses) and software tools at business users, and said "Go at it." That's simply been too hard. We used a lot of terms like "ad hoc queries" and "drill down," but it simply didn't happen very often.

I've argued for a while that organizations need to increase their focus on decision-making. In particular, they need to think again about the relationship between information and decision-making. I recently completed a study on this topic, with the sponsorship of IBM's Information Management business unit, in which I looked at 26 efforts to improve decision-making in organizations. I concluded the following ten things about how business intelligence (BI) needs to evolve:

1. Decisions are the unit of work to which BI initiatives should be applied.
2. Providing access to data and tools isn't enough if you want to ensure that decisions are actually improved.
3. If you're going to supply data to a decision-maker, it should be only what is needed to make the decision.
4. The relationship between information and decisions is a choice organizations can make--from "loosely coupled," which is what happens in traditional BI, to "automated," in which the decision is made through automation (see graphic below):

The Relationship Between Decisions and Information: Three Options

5. "Loosely coupled" decision and information relationships are efficient to provision with information (hence many decisions can be supported), but don't often lead to better decisions.
6. The most interesting relationship involves "structured human" decisions, in which human beings still make the final decision, but the specific information used to make the decision is made available to the decision-maker in some enhanced fashion.
7. You can't really determine the value of BI or data warehousing unless they're linked to a particular initiative to improve decision-making. Otherwise, you'll have no idea how the information and tools are being used.
8. The more closely you want to link information and decisions, the more specific you have to get in focusing on a particular decision.
9. Efforts to create "one version of the truth" are useful in creating better decisions, but you can spend a lot of time and money on that goal for uncertain return unless you are very focused on the decisions to be made as a result.
10. Business intelligence results will increasingly be achieved by IT solutions that are specific to particular industries and decisions within them.

There was a story in this week's InformationWeek that I believe strongly supports this set of ideas. One quote is particular perspicacious:

BI historically has been about dashboards and scorecards developed for specific uses, says AMR Research analyst John Haggerty. But that's changing. "All of a sudden it's about integrated analytics within applications," he says. "The conversation is starting to shift to looking at information in the context of specific decisions and roles."

My sense is that this shift will be a major one, and both vendors and users of BI will be transformed by it. What do you think?

The Wisdom of Geeks, The Madness of Crowds

2008-11-14T21:32:38Z

Warren Buffet recently commented, "Beware of geeks...bearing formulas" with regard to the financial crisis, but geeks bearing formulas did pretty well at predicting the results of the Presidential and Congressional elections. Better, even, than the much-praised Iowa Electronic Markets.

The geeks in question are those at 538, which is both the number of electoral votes and a relatively new political website run by Nate Silver. Silver was the subject of a highly laudatory New York Times article this week, and he deserved it. Not only were his final presidential popular vote predictions highly accurate (predictions of 52.3 percent for Obama vs. 52.4 actual, and 46.2 predicted for McCain vs. 46.3 actual), he predicted that Obama would beat McCain as early as last March. And while some Congressional elections are still not resolved, it appears that his predictions are going to be very accurate there as well; he correctly predicted all the resolved races in the Senate, for example.

Silver has, I believe, discovered a good method for political prediction. You start with polls, which are now both frequent and widely-dispersed across the land. You take a bunch of them, correct for biases of various types (e.g., consistent lean right or left, or pollsters that don't call mobile phones), and then run a Monte Carlo simulation on the outcome to see what the probabilities and confidence intervals are. As you can see if you look on the website, it results in a massive amount of data, but it seems to work.

The Iowa presidential prediction market did well too in the end, but had some crazy blips for McCain in late April and May. At the end it forecast Obama's popular vote almost perfectly (52.5 predicted vs. 52.4 actual), but was a bit high for McCain (47.9 predicted vs. 46.2 actual). Even if it were as accurate as 538, the prediction market is probably not well-suited to dealing with electoral votes for President or state votes for Congress. That's just too many things for people to bet on, and it would undoubtedly confuse bettors.

Although the Iowa market as done very well over the years, it's subject to the mass hysteria that we are seeing in financial markets right now. Charles Mackay didn't call it the "madness of crowds" for nothing. I can't really remember back that far, but I suspect in late April and May there was a lot of media speculation that McCain--already the presumptive nominee--was going to jump all over whoever the exhausted Democratic candidate happened to be. And Iowa bettors bought that logic, at least for a few days. The behavior of these bettors shows that the Iowa markets violate a key principle of Jim Surowiecki's in The Wisdom of Crowds: that the predictors are independent of each other.

The 538 approach is another approach to crowd wisdom--it's an average of polls with a lot of statistical massage thrown in. Averaging polls even without the quant manipulation can be pretty accurate--see the predictions, for example, in Pollyvote.com or RealClearPolitics.com. The aren't quite as accurate as 538, but they don't exhibit the fluctuations of the Iowa market.

One of the most interesting questions about all this may be: what's next? Nate Silver did great work in baseball analytics, developing the PECOTA system (Player Empirical Comparison and Optimization Test Algorithm) that's now a part of the Baseball Prospectus offering. Then he brought analytics to political polls. What new data-intensive domains remain to be conquered--by him or some other geek?