Data Quality: Getting Started

If you happen to be a "mover and shaker" or if you aspire to that role, then you'll be looking for access to one or more key decision makers of influencers closest to the top of your organization. You'll be determined to convince those people that

1. investing in data quality is a sound business decision
2. you are the right person to produce the ROI that they'll be looking for

If, however, you simply want to make things better as soon as possible and create new friends and allies while doing so, then you may want to take a different approach.

My recommendation is to use the tactics of the Special Forces. The "Green Berets" were formed into small teams comprising skills critical to the people they were trying to help. They then went out to those people and lived with them. Doing this allowed them to gain credibility and to learn what kinds of changes might (or might not) be acceptable.

The Green Berets helped the people with their work and, while doing so, offered improvements--small changes that produced higher productivity or more consistent results. The goal was to create allies.

"Data Quality" represents exactly the same kind of productivity and/or consistency improvement for our "indigenous" people in whatever part of the company they may serve. A DQ Team may be as small as one member and can produce results that are shocking in their scope and value as well as in their lack of cost. It isn't necessary to spend long periods of time "living" with the people. In fact, one lunch or a couple of coffee breaks will do IF you

1. ask the right questions
2. listen carefully to the responses
3. offer support
4. follow through

You'll ask about what happens when they get incomplete or incorrect forms (data is usually thought of collectively as a form) from their internal "customer". You'll ask about the extra work they have to do in such situations. Be prepared for an emotional response, this is what causes them to miss deadlines, work overtime, add staff... Also be prepared to hear that they simply pass the problems on because they aren't staffed to deal with them and don't feel accountable for fixing problems they didn't create.

Listening will uncover the sources of the most frequent or egregious DQ errors. Now you can mention that you are about to begin a project with those dirty so-and-so's, that it's likely they don't even know the hardships they're creating, and that you'll be happy to mediate a discussion amongst the parties to try to find a resolution. You may already have some ideas.

Create the meeting, making sure that ALL parties are represented (you were listening carefully, right?) and facillitate the discussion, if necessary gently guiding the discussion but never offering solutions. When the solution is "discovered", the people will own it and will implement it with minimal assistance from you. If your assistance is required, make certain that you deliver and do not hold them up.

Follow up by monitoring, coaching, facillitating and then ask if they'd like some help in publicizing their success. Because you know important people, they'll almostly certainly jump at this opportunity. You give them all the credit and they'll make sure to let everyone know that your help was both timely and critical.

This approach works and can even result in regular meetings to follow the improvement and to look for new opprtunities.

Two approaches--you choose which one has the highest probability of success for the greatest number of people in the shortest time at the lowest cost.

The Theory of Everything

The US economy, so far as the majority of citizens is concerned, is in the toilet and swirling rapidly in a clockwise direction. Health Care, long in its own toilet, has at least stopped swirling momentarily. All of us have a stake in what happens in those toilets.

I have a stake in another toilet as well. The portion of the economy devoted to technology has been caught in a vortex since the dot com implosion 15 years ago. I realize it will do no good to link all of these since linkage simply makes the resulting mess seem even more impervious to any corrective action.

However, (deep breath) if we don't consider the nature of the connection between them, we have very little chance of making sustainable progress in any of them. So, as my contribution to posterity, I nominate the accension of appearance to the pinnacle of importance in decision making as the criterion most likely to be acknowledged as the root cause of all three problems.

Since 1950 the rate at which appeareances have displaced substance as the motivation for decisions in the US has increased at a dizzying rate. In the past two years I have seen the nations physicians, as represented by a blue ribbon panel from the Mayo Clinic, state that the answer to the nation's health care woes is better access to insurance. The calling that was Medicine has emerged as a new entitlement program for the elite.

In the economy, fiduciary responsibility has been replaced by revenue numbers as the force that justifies all decisions.

In the Technology world, the means have come to justify the ends. Any decision can be justified if it allows me to position myself as a front-runner, new, hip, cool. "There's an app for that" allows us to spend unjustifiable amounts of money just to have that app in our pocket. Similary, corporations spend unconscionable sums on technology projects for which the need is poorly understood. Because the "solution" has to be new to give the proper appearance of tech supremacy, the outcome is always in doubt. Risk isn't so much managed as PR-ed. Spin control is the name of the game.

In almost 30 years of working with technology I have learned one lesson that transcends all others:

Either control your technology or it will control you.

In assessing the meaning of this for your own situation it is well to remember that

• Technology demands consistency
• Humans and human organizations are incapable of consistency

Just a few recommendations

• Be clear about WHY you want to do something
• Insist that others are clear about WHY they want to do something
• Choose a path that is known to produce the desired result--or at least choose next steps that are known to produce appropriate results.
• If you are unable or unwilling to do the above, choose another line of work if possible and stop complaining if not.

We are all either a part of the solution or a part of the problem. In either event, complaining about what someone else is doing or not doing will produce absolutely no change.

If you can't tell the difference between substantive value and the appearance of value you should avoid positions in which you will be called upon to make decisions. If you can tell the difference, then for all our sakes, make the decision and don't give it to someone else.

Christmas Wish List

The 13+ years I have spent in Healthcare have sensitized me to some things. I might have preferred to remain ignorant of many of them. In the spirit of Christmas, which is handy at this time of year, I'd like to nominate a list of gifts that would bless all residents and citizens of the U.S.A., regardless of theology or philosophy. Each item in the list relates to health care.

1. I wish that the role of technology could be clearly understood. There are a vast number of well-funded voices who want us to think that technology is health care or that health care is technology. In reality, technology is best thought of as a tool--an inert and often expensive piece of equipment, which, in the right hands, can produce wonderful results.
2. "The Patient" or "our patients" is not the same as "my patient" or "Josie Jones, patient". It may not be possible to apply technology designed for delivering care to a generic patient to Ms Jones. That doesn't mean that the technology is bad. It only means that the technology must allow for deviation in procedure. I wish that the role of abstraction is system design could be clearly understood.
3. I wish that all of the factions in the healthcare struggles were clear about their goals--with themselves and with each other. Only by being self-aware, open and open-minded can the parties negotiate a solution advantageous to all. Doctors, nurses, administrators, vendors, government and patients are currently at odds. The friction is not only between factions but within factions. Who will speak for physicians? The A.M.A.? Mayo Clinic? Who? Who speaks for government, for vendors, for patients, for nurses, for administrators? Each of these groups functions like a mob--surging to and fro as a strong voice emerges and then is drowned out. Each group must organize itself before "healthcare" can be organized.
4. Though I recognize that individuals and groups may be driven by ego to appear more knowledgeable than the next, I most devoutly wish that each of us might recognize that the person across the table might actually have some knowledge that we don't. I wish that we would listen first and assert only when necessary. I wish that we could see ourselves as occupying the same driverless bus.

There are many more things I might wish for this year but I don't want to seem greedy. May you each be showered with blessing upon blessing as one of God's beloved and may we fully appreciate each blessing as it comes.

Principles of Data Governance

Malcolm Chisolm, in a recent column in Information Management (A Principles-Based Approach to Data Governance) raises an excellent point. In 2006, when I attended my first Data Governance Conference, there was much discussion around a definition of DG. Implicit in this discussion was the need for something that was concise, yet comprehensive, and on top of that, engaging. The idea was that this definition could be used:

• As part of a sales pitch (like a slogan)
  
• To create synergy within the emerging discipline
  
• To provide focus for any ongoing methodology efforts

Some present may have had additional motivations, but I think these were the ones on most people’s minds.

The definition that emerged was acknowledged to be a work in progress. By the 2008 Conference, one of the tutorials quoted three definitions:

1. Data Governance refers to the organizational bodies, rules, decision rights and accountabilities of people and information systems as they perform information-related processes.


2. Data Governance is the practice of organizing and implementing policies, procedures and standards for the effective use of an organization’s structured/unstructured information assets.


3. Data Governance: The execution and enforcement of authority over the management of data assets and the performance of data functions.

These were troublesome to me then, probably for the very reason that Malcolm mentions. All seem to acknowledge a context based on an organization’s information assets, but their focus seems to be quite different. The feeling I have is that they are advocating a judicial, legislative and executive approach to governance.

In the U.S., a Constitution lays out these three perspectives and establishes the mechanics (framework, architecture) within which governance will be administered. Within the Constitution and before any of the mechanical parts are discussed, in fact, within the preamble, first principles are asserted. The writers tell us that what follows will be a system of governance for the purpose of

• Forming a more perfect union


• Establishing justice


• Insuring domestic tranquility


• Providing for the common defense


• Promoting the general welfare


• Securing the blessings of liberty to ourselves and our posterity

While this model would probably work on its own in establishing [data] governance, there are just a couple of nuances that will have to be accommodated because our system will not be working in a representative democracy but in a corporation.

Within the context of our system, leaders are appointed and serve at the pleasure of stockholders rather than the public. The principle of one-man-one-vote does not apply. One person may control sufficient votes to dictate to the Board of Directors. Within the day to day operations, the ability to dictate policy, direct activities and appoint deputies is granted at multiple levels, though always subject to the pleasure of the higher levels.

Having now established a context, it’s time to agree on some first principles for data governance. The candidates are:

• The entire corporation must agree to be subject to the system.
  While those placed higher may still, at their pleasure, appoint and dismiss deputies, they must agree that [data] governance operations will be a factor in those actions.


• It must be understood that within the corporation, domestic tranquility, the common defense and the general welfare are all dependent upon the information assets owned and managed by the corporation.
  When the system is followed, all processes will flow smoothly, problems are addressed at the process level, personal antipathies are secondary to process execution and process anomalies such as unplanned rework and delays are greatly reduced in number or even eliminated completely.


• Consistency is everyone’s goal.
  In a work context, surprises are almost always seen as negatives. Our goal must be to improve the consistency of our processes and their outputs such that surprises become exceedingly rare (six sigma has been suggested as a goal) and predictability becomes commonplace.

These principles should be the touchstone(s) of our efforts. Everything we do should be evaluated on the degree to which these principles are addressed.

I will suggest that these may also be the principles of the corporate Quality Assurance effort and remind everyone that they are also the basis of Deming’s 14 points as well as other quality improvement methods. No improvement is possible without first establishing a stable (consistent) process.

I leave you with one final principle: Data governance will not be implemented as a stand-alone initiative. If we cannot see data governance as part of a larger, comprehensive system of governance, we will not be able to address any of the three principles suggested above.

Survival, Error & Technology

I'm going to pass on some wisdom here. It's not very often that we encounter wisdom today, especially where technology is concerned, and it's often the case that we don't recognize or acknowledge wisdom until we're looking back over the wreckage and trying to figure out what we should have done. I'm probably also setting myself up by labeling this as wisdom but I am so weary of seeing the same ads with different acronyms and talking to the same people with different names.

You will never find your way out of the current mess you're in or about to be in by searching for and hiring someone with recent experience on a specific product. To put it another way, a specific product, no matter how much buzz it enjoys, is never the answer.

I will be among the first to acknowledge that the use of absolute language (never, always...) and even the use of unqualified superlatives (best, worst, fastest...) is a habit to be avoided, nevertheless, decades of experience have proven that the absolute statements in the preceding paragraph represent wisdom and that failure to heed this wisdom will produce cost overruns, timeline disasters, confusion, stress, employee turnover and a host of other undesirable outcomes.

In large part, the success of the human race has been due to our ability to recognize exceptions without necessarily understanding the rule. My own take on this is that, with today's reliance on technology, we may have reached the point where the process of natural selection that has honed this skill over countless generations has now produced a liability. "Something's different," is enough to put us on guard and may be enough to launch a complex defensive reaction to preserve the safety of the individual or group.

First of all, while it is still good to recognize exceptions, it is now absolutely essential (that's an absolute absolute) that we develop the ability to recognize the underlying rule. A study of human error (Human Error, Set Phasers on Stun...) shows that leaping to conclusions about the rule is what produces the error condition. In fact, if we can't describe the rule in terms of the logic of the computer (if... then... else...), we can't rely on technology at all.

You might ask, as I did, how we might acquire this ability. The time tested way is known as [survivable] experience. There are a host of cause-and-effect analysis tools and techniques that have the appearance of rigor and reliability and are an improvement over experience, especially when combined with exhaustive testing, but you will find that even these are more productive when used by people with experience in the world being analyzed.

Tools are great and another critical human enabler, but--and this can't be over-emphasized--no tool is so advanced that it runs itself. Every tool, no matter how advanced the technology requires human hands and a human mind to guide it. If you were to be presented with the greatest woodworking tool in the world or the most advanced sewing machine or fishing gear or computer, would you immediately become a master cabinetmaker or designer or fisherman or software developer? You might note that the only immediate change is one of expectation.

An experienced person with rudimentary tools is more likely to produce a quality result than the inexperienced person using the "best" tools. The fact that I used a tool just yesterday says nothing whatsoever about the level of my experience in producing the required outcome. I have made the mistake of looking for help and focusing too narrowly on what amounts to recent experience with the tools in my shop. I have learned (through experience) that I will enjoy better results if I'm learning while interviewing my prospective employee. If I'm talking with someone whose knowledge stops at the tool's user interface, then I had better be prepared to devote myself to directing the employee's work. If I have a staff composed of such employees, then I need to possess all of the requisite experience myself or else be prepared to conduct a project whose principle product is more experienced workers.

The challenge is to find the right mix of experienced people in supervisory or team lead roles and people who possess dexterity but are in need of experience. If I'm in a director or management role, I have to have experience producing a product with that scope. A technology "system" has a complexity that is beyond human comprehension. The only way to design and build it is through a process of identifying smaller and simpler pieces, building those and then assembling them into the final product. You need to look for people who have an appreciation for the amount of effort this takes and the discipline--both personal and organizational--that it takes.

Stop looking for Oracle or CRM or Rational or even "use case" or "data model" experience except as clues about the approach that the candidate might be expected to take. I understand that these things are ideal as targets for a logic rule processor, but the rule ("find resumes that include these terms") is so simple-minded as to be useless. If your only goal is to turn 1000 resumes into 100, then proceed, but if your goal is to find someone who can get you out of the predicament you're in, then you should spend more time on your rules so that the exceptions are more productive.

Business, Information and Technology

Are you in management? Do you have annual/quarterly goals? Will you be held accountable for achieving those goals? Is the accountability expressed in bonus dollars? Is there any possibility of a zero bonus?

Are you still with me?

How will the achievement of your goal(s) be measured? Please note here that "how" has two dimensions: one is related to process and the other to a unit of measure. In all of my vast personal experience, all of the attention has been focused on the unit of measure part (when there has been any attention at all), and the process part has never even been part of the conversation.

Please understand that what follows is not intended to sling mud at any individual or organization. My purpose is to clear the air so that we can talk about how we're really going to achieve our goal(s).

I am going to generalize based on extensive, though anecdotal, experience. In other words, I have not conducted a survey, scientific or otherwise, and cannot produce data to back up anything I'm about to say here, so I'm leaving it up to you, the reader to determine whether it feels like truth or not. Should you feel that this does not ring true, or should you wish to fault me for not being more objective, I would ask that you produce a study or at least a body of experience in support of your position.

Awards of bonus dollars tied to achievement are based solely on whether the holder of the dollars wishes to give them away or not. There is rarely, if ever, any protocol defined for defining metrics, units of measure or measurement process. You will go into an "annual assessment" meeting with your boss and he or she may discuss your level of achievement in very general terms before announcing the amount of your bonus or a recommendation for an increased level of compensation.

Why is this important, you ask? Well, it is important because it's the the way things are done. Despite vigorous protests to the contrary, the business world is set up to run on subjective assessment supporting subjective decisions. What, you say that your decisions are "data-driven" (objective)? I would love to hear the story behind the data that was used to arrive at your most recent decision.

All of this is background for understanding why "data" initiatives so frequently become mired in a swamp of politics and personality. Let's walk back from a data-driven decision.

1. You are able to make the decision because you trust the data.
2. You trust the data because you are familiar with and trust its source.
3. You trust the source because you know that it is reliable.
4. You know that it is reliable because it consistently produces information that can be relied upon.
5. The source has been consistent because it always uses a tried and true methodology (set of processes) to produce its product.
6. The consistency is possible because the methodology includes steps designed to validate the source's inputs.
7. The validation decision returns us to #1.

How do you feel about standards (e.g., standard operating procedure or SOP)? If you don't currently support the creation, implementation auditable use of standards or, at any time in the past have not done so, you have no right to and almost certainly do not have access to reliable information and therefore no claim to data-driven decisions. Just to drive the point home, when your boss decides that you won't be getting that bonus or increase you were counting on, your only acceptable response is to smile and say thank you.

By the way, if you notice that your bank account (or budget) is suddenly much bigger (or smaller) than it was yesterday, what is your responsibility? Who are you accountable to? How much trust can you afford? Now you have some insight into compliance.

The use of technology introduces an additional huge portion of uncertainty into the trust equation. Take another look at the decision walk-back above and note the points where the use of technology means adding additional paths and complexity to the validation process. This is what your data governance people are trying to get their arms around.

To summarize:

• data-driven or intelligence-driven decisions demand trust
• trust demands reliability
• reliability demands consistency
• consistency demands compliance
• compliance demands governance

OK, you can go back to work now.

What Is "Data" Anyway?

If you have any experience with phone support (on either end) you will recognize how easy it is to get deep into a process before realizing that the other person is on a completely different path than you are. RTFM (Read The F---ing Manual) often pops up as the answer to our communication difficulties, but it clearly is not the answer or it would have been universally embraced by now.

I happen to be a proponent of the theory that the answers aren't nearly as important as the questions. As a teacher, I know that learning is happening when the pupil is asking questions--particularly a series of related questions. This has become very important as I have attempted to make headway on [data] governance and [data] quality.

My early attempts assumed that everyone knows what data is--and they do, in the same way that a picture is worth 10,000 words. Each and every person you talk to knows what "data" is and each has a different idea in mind. For most, it's a picture of the last set of values they looked at. This might have been a spreadsheet, a graph, a collection of measurements... The key is that data is a set of values. For some, "data" is a commodity. It is files, stripes on a disk, a percent of capacity, a quantity of bytes measured in "mega-", "tera-" or "peta-". For still others, "data" is represented by a schema, model, definition, or some other abstraction.

Given those varying perceptions or perspectives, is it any wonder that at some point in the quest for "data" anything we find ourselves stuck in the quicksand of confusion. Even when all parties have been saying the same things and have been agreeing on goals, there comes a point when someone will say, "We're not going to do that." or "I don't see why that is necessary." or "But that will change my work flow." This is frequently the point at which everything starts to unravel.

So what I have learned, and what I offer to you now, is that the initial phase of any data initiative must be a carefully constructed education process to insure that the quicksand moment never happens. This must be thought of as risk management. Remember, too, that the really important questions (and answers) initially are not the ones you're hearing. All of the different constituencies are going to be much more comfortable exchanging information (or misinformation) within the tribal group than with "outsiders."

The best way to head off this risk is to carefully choose allies from each constituent tribe and use informal conversation about their pain points and the ways that data figures in the relief of that pain. You will be setting these people up to be the "experts" within their respective tribes. Part of this will be coaching them in how to respond to questions and discussion in which they don't feel themselves to be on firm ground. They need ways to postpone a response until they've had a chance to confer with other experts. This is easy to do by setting up a collaborative model.

Rule one of this model is that I never answer for someone else. Everyone can understand that a situation involves yet another perspective and that it is necessary to involve someone from that tribe in order to get a complete answer. The most common danger here is "We don't have time for that." Everyone must understand that this is an absolute red flag event. It signals that we still have not achieved a universal understanding of objective.

When a red flag event happens, it isn't the same as finding ourselves up to our necks in quicksand. It just means that we need to engage in some risk mitigation. It's a sign saying "Quicksand ahead." We will need to bring this person into the fold--usually through informal and non-threatening discussion with at least one peer or trusted expert.

Your role, should you choose to accept it, is to be a non-judgmental, constant, committed, and helpful presence that can be relied upon to be a neutral mediator and facilitator who feels like a friend in any need. Your motives must be above reproach. You cannot count on and should not hope for recognition. All around you will be better off for your presence.

If you are senior in the organization to this person, you should make sure that you are appreciating their contribution but they will appreciate non-public affirmation since putting them in a spotlight may negatively impact their ability to continue to function in the same way.

"What is data anyway?" is a question that requires many answers initially and one answer eventually. Remember, though, that many people are really only interested in what they have to do differently. "Data" may have no meaning whatsoever in their day-to-day responsibilities even though they may be monitoring real-time run charts with instructions to take a specified action when the line goes above or below a certain point. You can't possibly know where to start or where to stop in defining data for them. That's why you need the tribal expert.

Don't seek "important." "Helpful" will take you much farther more quickly.