Quality management in a learning organization

I mentioned before that the American Society for Quality held their World Conference on Quality and Improvement last month. As with any such conference, there were keynote speakers and then multiple breakout sessions to present special topics. One of the side-effects of this format was that different topics built on each other in unexpected ways. 

And so early on the second day I found myself wondering, We talk a lot about "learning organizations." If we want to create such an organization, how should we design the management system in order to support it? I don't claim to have an answer that is final or complete, but I started to collect some notes. Also, I hope you will join the discussion with ideas of your own. With luck, your ideas will be better than mine, and I will learn something.

I said this topic occurred to me early on the second day. The inputs from the first day, that started me thinking in this direction, included the following:

• Ugwem Eneyo, CEO of Shyft, discussing in her keynote address how technical development in Africa is leapfrogging over the stages familiar in Europe and North America to move in new and unexpected directions
• Charles Cox, on "QMS for data driven decision making and growth"
• Jane Keathley, on "Aligning organizational structure with Quality 4.0"
• Jeremiah Genest, on "Enabling the process owner to drive improvement"
• Heather McGowan opened the second day with a keynote on the rapid rate of change in the modern economy.  

And my initial thoughts, in no particular order, included the following.

A Quality Management System includes the following elements:

• analysis of the context of the organization
• policies, procedures, and records
• roles and responsibilities
• competence management
• tools for problem-solving
• internal audit
• management review

Regarding documentation and records, Cox made the point that, if you don't know what you need, all documentation systems look equally good. But in fact, different systems are good for different things. A learning organization needs the ability to create testable hypotheses and then record the results correlated with those tests, in order to evaluate the results for future planning. In addition, a learning organization has to get feedback as soon as possible; so whenever possible, long-lagging indicators should be replaced with short-lagging ones, and short-lagging indicators should be replaced with leading ones. It's not always clear how to define KPIs in this way; but whenever it's possible, it should be done.

If the organization truly is learning how to do things in new and better ways, roles and responsibilities are unlikely to remain fixed. Keathley described a networked organizational structure as one in which organizational charts are temporary, designed around particular initiatives; and such a structure sounds like it should be appropriate for a learning organization. There may be constraints imposed by the technology you have to deploy: the structure in a manufacturing plant is likely less fluid than that in a service organization, because using the machines correctly requires specific technical skills that not everyone shares. But again, so far as a networked structure is possible, it is probably the way to go.

In a learning organization, there is always the risk that a specific set of skills, or a specific job classification, might become obsolete—a risk that McGowan explained is inherent in the modern economy anyway. In this context, competence management means more than just training people to do a job. Rather, for each role the organization should identify the needed competence profiles, and then forecast what jobs with similar profiles are likely to be needed in the future. The people working in today's jobs should then be offered continual learning opportunities to expand the number of useful roles they can fill. 

By the way, it is critical that these opportunities be communicated in the right way. The message must never be, Your job is going away so you better learn something else instead. In the first place, that message is dispiriting; in the second place, it could be false. A "learning organization" is one that doesn't know the future yet; so the expectations for the future might not come to pass. A better message is, You are already good at A; we think that in the future we might need someone who knows how to do B, which is similar; and you already know us and our organization, so you would be a natural for this new role if it opens up. Would you like to learn more about it? Then even if it turns out that you don't replace role A with B, your organization will be stronger for investing in employee competence. And your employees will know that they get something out of it too.

In a learning organization, internal audits take on a new role because they are no longer simply tools for enforcement. Yes, of course an auditor has to check whether the defined procedures are being followed. But the auditor also has to check what results the organization is achieving. If the procedures are not strictly followed but the results are still good (like in my story half a year ago about the plating bath), this may be evidence that the organization can rewrite the procedures to make them looser or more generous, or to allow alternative ways to get the job done. If the procedures are followed perfectly but the results are poor, this too may be evidence that it's time to rewrite the procedures, because obviously they are no longer effective. Either way, the organization has to treat the audit as a fact-finding event, and then has to evaluate what the information really means.

Management review may change most of all. During management review, you check all your process metrics, to assess how things have gone in the last quarter or the last year. But in a learning organization, the point—oddly enough—isn't to meet your targets. The point is to learn how to do better. Targets in a learning organization are set at least partly as scientific experiments, to test hypotheses. If the numbers come back green, that shows the hypotheses were correct: but that's not necessarily a good thing. It depends where you want to go next. On the other hand, if the numbers come back red, that shows the hypotheses were incorrect: but again, that's not necessarily a bad thing. Even if your first guess turned out wrong, you now have more information than you had before. You have learned something. And you can do better next year. 

There's another element to learning, too, and management review is where it happens. Peter Drucker used to make the point that companies must regularly re-think what business they are really in, because the business landscape is always changing. Who are your real customers? What do those customers really want? These answers are never fixed, but too many companies treat them as immutable truths, laws of nature that never have to be reconsidered ... until one day the company goes broke because its old market ceased to exist while nobody was looking. It is exactly these questions to which a learning organization must learn and then relearn the answers, evaluating all the data from the whole Quality Management System in order to keep a step ahead of obsolescence. But to do this, top management has to use management review to force different questions than it asks in most companies. And it can never settle for easy answers.

All this is just a sketch, just a beginning. There is far more to work out. So now it's your turn. Leave me comments on this post to tell me where I'm wrong, and what I've missed. Figuring this out will be a collaborative effort, and I know I need your help.

What is more, our organizations need to know the answers so they can become better and more resilient. So please, give me your feedback.