Saturday, March 29, 2008

Last night I attended the University of Chicago GSB Consulting Roundtable for a very informative presentation on Web 2.0. One of the applications we talked about was using the social network structure to do ideation. The process starts with a straw man idea. This initial idea is sent to two people who add to, or modify the idea -- change it to create a somewhat new idea. Each of these people sends their idea to two additional people who modify it, send it to two other people, and so on. Each participant gets to rate and rank the ideas, and ultimately there’s an idea(s) selection process.

It will be interesting to watch how this ideation process pans-out. It is a popular belief in innovation circles that to get a really good idea, you need lots of ideas. This is the old Edisonian approach to innovation. There is nothing wrong with generating lots of ideas. If you’ve got to solve a problem, and you don’t have a better method, then lots of ideas is a pretty good approach. You’re likely to generate a reasonably workable idea.

I think there is at least one great benefit to this way of ideation. In any business there is a chance that a strong synthesizer is close to the problem. It’s highly probable that this synthesizer has already thought through the problem and has a good solution. In many businesses, there is no good way for this kind of person to surface an idea. This type of ideation process may allow the idea to be surfaced as a ‘modification’ of the straw man. It still may go nowhere because often the change that’s needed cuts pretty deep into what’s in place.

People are naturally cautious about big changes. Part of that is because big change usually carries a degree of risk. But it is also true that our brains form increasingly rigid ‘success’ patterns around processes that have repeatedly worked well in the past. These hard wired patterns blind us when we examine a new idea. It is why in hindsight a hard fought for solution looks obvious when the dust settles. Oddly enough, the best ideas are almost always initially rejected with statements like: “the idea is completely unworkable”; “he doesn’t know what he’s talking about”; or “we already tried something like that and it was a total failure”. People’s brains reject change to a process or product that has long been successful. The real obstacle to this Grapevine ideation process is the notion of rating-and-ranking ideas. It is a popularity contest. The best idea will likely be pretty unpopular initially. What is needed is problem-solution evaluation criteria.

Lots if important innovations did not come from some group-think (after all, each act of synthesis happens in just one brain). The literature is full of examples of important innovations that were rejected repeatedly as dumb ideas. It is just how our minds work. Thomas Watson of IBM initially rejected the computer, 3M repeatedly rejected masking tape, Xerox rejected the personal laser printer (which HP then made into a very profitable business), and HP also rejected the display monitor, etc. It was the persistence of the originator that eventually turned the idea into a delivered solution.
One area where it would be interesting to test the usefulness of the Grapevine approach, would be in the initial problem-definition stage. That is, use the tool to define the problem precisely, with the rate and rank process identifying the fundamental properties of the problem, then the solution can be evaluated in terms of how well it purports to address each of these properties.

Wednesday, March 26, 2008

Knowledge in Business

Browsing on websites about ‘Knowledge’ I came across a site that had this statement prominently displayed:

Ultimately, knowledge in business only has value if it results in action.

At first this looks a bit like a ‘mom and apple pie’ kind of statement. But when I though about it, I realized that the statement is just wrong. This suggests that if a company is going to invest in knowledge, then the company ought to reasonably expect the knowledge will result in action. By implication, the company would not invest in knowledge, if that knowledge did not have an obvious potential to result in action. Hence the engineering department would pay for engineering classes, but not learning to signing for the deaf (assuming no obvious link to action.) Someone who does not understand the role knowledge plays in innovation must have made this statement.

Synthesis, the combining of existing knowledge to form new knowledge, is the primary source of innovation for most companies today. Very often the solution to a problem comes in the form of a metaphor from some distant knowledge. E.g. the ping-pong ‘burp’ gun leading to the ‘sinking’ pipeline solution (described in another blog, [Example: Synthesis and the Obvious Solution] below). We are not able to predict what knowledge will provide a pattern, which drives our brains toward a particular solution. Hence, broad knowledge, with lots of ‘waste’ is better than narrow knowledge which does not give our brains the distant metaphors we use to solve problems.
The second reason that obtaining knowledge in an alien field is valuable, is what happens to our brains as we age. The metaphor of a rubber band works well here. If you do not take care of a rubber band it loses it elasticity and, after a time, it will snap. However, if you take care of a rubber band, you can maintain its elasticity. The human brain is the same way. Our ability to see things differently, to find new patterns from our knowledge depends upon the elasticity of our brains. Neurologist, like those at UC Berkeley have shown that if we keep challenging our brain to learn new things, things that are very different from what we already know, then we can maintain our brain’s elasticity. Otherwise, it loses its elasticity and our ability to see things differently ossifies. That is, if you speak Spanish, don’t choose Portuguese to challenge your brain, it is too similar to Spanish. Instead, learn to sign for the deaf, or read Braille, or learn to sculpt or study Chinese literature. Choose something very different from anything you’ve been exposed to before.

Friday, March 21, 2008

Is Experimentation Innovative?

Experimentation requires diligence, rigor, meticulous adherence to technical methods and perseverance. In a corporate environment, good experimenters are people who are diligent and rigorous, and can follow prescribed methods without variance and will keep at it. These people do not get easily sidetracked.

Synthesizers, on the other hand, are constantly looking for new problems and newer ways to solve them. They are easily sidetracked, and would be more likely to perform each experiment a bit differently in an effort to produce a quicker, cheaper, more efficient result. They are also more likely to drop the experiment if it does not produce results quickly.

Experimentation is expensive, time-consuming, and does not guarantee a solution to the problem at hand. Companies that must experiment, like chemical and drug companies, are constantly looking for ways to shave costs and delay from Experimentation. That’s where super-synthesizers add real value. They see problems from different perspectives.

Such companies should seek out such synthesizer talent and incorporate it into the Experimentation process। They will find unseen problems and solve them with great expertise. Occasionally, a failed experiment yields a serendipitous discovery. A synthesizer will see a failed experiment from a completely different perspective giving the company a double opportunity to enjoy new IP from a serendipitous discovery.

Yes, Experimentation is innovative!

Sunday, March 16, 2008

Two tenets of good MIS

Two tenets of good MIS design are: 1) the information should be unbiased, and 2) the benefits to the firm of having the information should outweigh its costs. MIS that is designed to demonstrate the value to the firm of a department’s activity almost always violates both these tenets. When they produce MIS designed to demonstrate the impact on profit of their activities, it’s easy for support managers to rationalize that the MIS is cost effective and free of bias. But support activities such as those performed by building maintenance, the transportation pool, the janitorial services, etc.,. are done because they are deemed necessary to the operation of the business. They are already considered valuable and trying to put a profit measure around them will not alter that.

A janitor once tried to show the impact his work had on the bottom line of the firm. He undertook a study of the impact on employees when he failed to fill the men’s room toilet paper dispensers. Measuring the time delays caused by this lack of paper he determined that on average this would cost the firm 59 second per employee per day. He further then estimated the impact of filling soap dispensers, vacuuming the carpets, sweeping and mopping, emptying trash cans, etc.,. Ultimately his analysis showed that his work contributed close to $1.5 million per year to the firm’s bottom line.

Even if we assume the janitor’s information is unbiased and accurate। What is the value of this information? Is the firm going to go out and hire 20 more janitors and realize a $30 million benefit to the bottom line? Is the firm going to give the janitor a ten-fold increase in salary? Is management going to look at the janitor’s activities with a new sense of awe?

When a department mixes useful management information with MIS of this nature, it undermines the perceived value of the information. It is important for any support function to be careful to only produce information which will have a meaningful impact on decisions which affect the bottom line.

Thursday, March 13, 2008

Adults Manage Themselves

My first line job was as supervisor for the "back-end process" in a Chargeback unit of the nation's largest merchant processing bank. Chargebacks are disputes in the credit card business. They are initiated by the bank, which issues cards to card members. The dispute can be a customer complaint, a fraud case, or a processing error such as a duplicate charge,. The issuing bank sends the dispute to the merchant bank using a Chargeback transaction. The merchant bank has a fixed period of time, depending on the nature of the charge back, to either remedy the dispute or charge it back to the merchant.

In the Chargeback unit where I worked incoming Chargebacks were distributed into two work teams depending on type of Chargeback. Typically, a ‘documentation’ Chargeback was more complicated to work than a ‘non-doc’ Chargeback. Every day the supervisors in each of these two departments would distribute piles of charge backs files to each of the Chargeback clerks. Chargeback clerks but were measured on the number of Chargebacks they processed each day. At the end of the day, any unworked files were returned to the supervisor. Once a Chargeback was processed it was sent to the quality unit, which checked the correctness of the processing before sending it either back to the issuer or onto the merchant.

In addition, all of the accounting control was handled in the back end where I was the supervisor. And the Arbitration unit handled all of the disputes that went to the card associations for arbitration. Altogether the Chargeback department consisted of 120 people. The unit worked a normal 40-hour week, and every Saturday about 25% of the people were required to come in to do overtime work. Morale was low. The unit was in a constant state of backlog, and because of the backlog timeframes were expiring resulting in losses for the unit. When I arrived losses were running at eight million dollars per year. Three people in the unit produced volumes of MIS around these losses.

During the first three months I was there I made some changes to the way the back-end operated to give us more control, and to automate the tracking of files by affixing a bar code to each. We installed three PC workstations with bar code readers so we could track the accounting slips associated with the files. I also spent time trying to understand the Card Associations’ rules around Chargebacks and the processes we used to work them in the rest of the department.

It seemed to me that the loss in the Chargeback department was due to four things. First, training across the board was inadequate. Second the organization of the work promoted inefficiency. Third, the way the performance was measured contributed significantly to the backlog. And lastly, the Quality Unit created it’s own backlog which allowed Chargeback timeframes to expire.

The supervisory team, met daily with the manager of the department. Most of the focus was on the MIS around losses. Losses were “the BIG problem for everyone on the team”. This MIS cut losses into every imaginable category, from issuing bank, to merchant type, to Chargeback type, to amount buckets, to aging buckets. None of the MIS helped identify the causes of the loss. On several occasions I attempted to make suggestions on how we could correct the four areas, which I felt we needed to fix. The manager responded on each occasion that I was new to the group, and that I should focus on my own area and give myself some time to understand the workings of the department before I made such suggestions. At some point, one of my fellow supervisors mentioned to one of his friends in another unit that I had ideas that would fix the problems, the friend mentioned it to his boss who mentioned it to his boss until the manager of the Operations Center got wind of it.

The center manager called me and we had a chat and he asked me to put together a presentation on what I would do to fix the unit. Ultimately, I made the presentation to the SVP responsible for the credit card business in the bank. The SVP gave me the go ahead. He decided to reassign the current manager of the department, putting me temporarily in charge.

I also exacted from him, a promise that we would not lay-off anyone in the department who was performing at a ‘level one’ or ‘level two’ performance grade, and that for six months no auditors would show up to distract us. Anyone from management was welcome to come and observe or participate unannounced.

It was mid-May, just when universities were concluding their spring semester, so I went to the nearby university and a hired six science students as summer interns to help with the department's organizational needs.

I called a general meeting of everyone in the department, and announced the reorganization. I told them that at the end of the year we would have 60 people working in the Chargeback department rather than 120, but that no one who is performing at a level one or level two would be laid-off, and that I would do everything I could to ensure that each person could demonstrate a level one or level two performance. Furthermore I said, if I have to lay anyone off, I will lay myself off first. The first requirement was that everyone on the staff get trained in the credit card rules, particularly around Chargebacks. We put together a simple, but effective, self-paced training program, which required passing a 25 question, multiple choice open-book test with a score of 100%. The test could be taken as many times as needed to get a 100% score. Each time the test was taken it was different. At first, the staff argued that the requirement was unrealistic, and perhaps a passing score of 70% was more reasonable. But, as people began passing the test, the objection disappeared. Several people adopted a learning strategy of just taking the test. We did not track how often anyone took the test, since in our view, the more often someone took the test, the better trained they would be. Within three weeks the entire staff had passed the test.

At the time, there were about 75 different Chargeback reasons. We assigned individuals to specific Chargeback reasons. These Chargeback experts were held accountable for ensuring that all the Chargebacks that flowed through their inventories were properly processed. We eliminated the quality unit entirely.

Since the volume of incoming work was completely out of our control, and it would be possible for some Chargeback experts to get ahead of their work and others fall behind, we established a performance measure, which encouraged teamwork. It was a simple structure which said you got credit for all the work done on the Chargeback reasons for which you were responsible (irrespective of who actually did the work), plus you got credit if you helped someone else to manage their inventory.

Human resources initially objected to the structure because it allowed us to give credit to two people for one person's work. We managed to convince HR that while this was true, the behavior would benefit us in the long run because, not only would we have a sense of team, but we would have overlapped in expertise which would cover us when someone was sick or on vacation. People are intelligent and clever, and before long we had people switching desks to get double credit. It cost us nothing, the quality of work was very high, and morale rose dramatically. HR further objected however, that too many people might achieve an outstanding performance rating. We asked them "what would represent a outstanding performance for the department as a whole?", and suggested that eliminating the losses, eliminating the overtime, cutting most of the fees we were paying for arbitration, and doing so with half the people surely would represent an outstanding performance. And if the whole group were needed to achieve this, surely there would be many outstanding performers. To which the head of HR laughed with obvious skepticism "if you could achieve that, then it would be 'outstanding'!"

Over a weekend, the management team along with the Science students reorganized the physical workspace and colorfully decorated the area. I turned the manager's office into a lunchroom, and moved myself out to a desk on a side of the floor where I would be accessible and everyone could see me. On Monday morning the trained workforce showed up to their new responsibilities. Overtime was immediately stopped. Within three weeks the backlog was eliminated. After six weeks all losses in the pipeline were flushed through, and further losses disappeared completely. So did the volumes of MIS around losses. Over the next three months, two people retired, one was fired because he refused help for an ongoing drug addiction, and 58 people landed other jobs within the bank mostly because the training we gave them made them the most knowledgeable people in the center. Among the 58 who left, were our 5 "best" people, all of whom received promotions in their move. We could afford to let them go because other members of the team were able to step in and be effective in those jobs.

Several months later we had a discussion about what made the reorganization so successful. I thought about that and said, "We took angry, demoralized people who were being treated as children and let them behave as adults." Their environment changed from having someone hand them their work, and having someone check everything they did, to an environment where they, themselves, managed the inventory, and no one checked their work. I was asked "If you don't check their work, then how to you manage them?" I responded, "I don't think you can manage adults. You can coach, help, and support them, but the whole notion that you need to manage adults is wrong. Adults manage themselves." We put together a structure in which people could understand how what was valuable and important to the business, was in sync with their personal and social values. Our success was based on candor, credibility, and the notion that everyone is unique, but, in our ‘adultness’ we are equal. Management's one responsibility is to manage the environment, the maturity level in the workplace.

Monday, March 10, 2008

Supersynthesizers and Solving Problems

One skill, which I think is the single most valuable element in finding solutions, is the ability to completely define the problem. This is a skill useful in both synthesizing through the logical analysis method, and synthesis that’s “flash in the sky serendipity”. Supersynthesizers have this ability and inclination to see the problem from different angles. For example, take the puzzle below.
Puzzle: This figure is made up of matchsticks laid out to form these five squares. Reposition two (and only two) matches to form four equal squares each with the same size and shape as the individual original five squares. You may not overlap one match on top of the other.

Most of us would tackle this by first looking for a solution. We ask ourselves – “Where might we be able to remove a match that would eliminate a box?” Then, we’d figure out where to use that match. We might do a lot of mental trial-and-error, and with enough persistence, either solve the puzzle or go crazy. If you are interested in doing this on your own, stop reading here and come back when you’re crazy.

But a super synthesizer sees patterns in problems differently from the rest of us. Instead of looking for a solution, a supersynthesizer might look at the problem and see if there is a different perspective. Instead of saying “I’ll start by picking up two matches” the supersynthesizer may say something like this. “I have to place two matches here such that they will each be a side to a box. I cannot place these two matches were a match exists today, because I would either be overlaying a match or just putting a match back where it was. Thus, I need to find a place where I can place two matches and form a box that is not already formed.” Or the synthesizer might say “I have 16 matches,” (how many of us would have even thought to count the matches), “and I need to make four boxes. Each box has four sides, 4X4=16, therefore, no squares can share a match.” Or say “I cannot take one or two matches from the bottom box with out leaving a match just hanging out, the same is true of the top box.” (although the logic for the top box is a bit more complex). “Thus, the adding and subtracting of the two matches has to be within space occupied by the other three squares.”

No solution has been attempted up to this point in the thinking, but the problem has been substantially re-defined. This makes the solution easier to divine. In a business context, the obvious solution often isn’t apparent until someone comes along and redefines the problem.