One of the questions I get asked often by my clients is “just how many KPI’s and metrics are enough for effective Performance Management to occur?”

I think most of them realize that there is no hard and fast rule, and probably no “right answer”. Some may just be trying to find out if they are “in the right ballpark”. But I think the majority of those asking this question are asking it rhetorically. That is, they believe “in their gut” that their measurement system has gotten a bit unwieldy, and is starting to create breakdowns, confusion and loss of whatever  momentum they once had. And based on my experience, they’re usually right.

Are we talking KPI’s or Metrics?

Now some of the Performance Management “purists” out there might say that we first must know what we’re  talking about when we say the words ‘too many measures’. Are we talking about KPI’s or performance metrics? Are we talking “business” metrics” or “operating metrics“? Are we talking about “data elements” (inputs), or are we talking about the calculated metrics that show up on our dashboards and scorecards that utilize these data in their algorithms? Or are we simply talking about operating data and variances that we routinely report and track in our budgeting and forecasting environments?

For today’s purpose, I’m going to stay clear of those distinctions. Mainly, because having too many measures can cause problems regardless of the type of measure it is. But also  because we’ve had more than enough posts this week (including mine)  that have discussed and debated what I would call “the semantics” or “the lexicon” of EPM. (Dashboards versus Scorecards, Enterprise vs. Portfolio Performance Management, Business Performance Management vs. the historical HR view of Performance Management (which interestingly suggests that HR is trying to manage something OTHER than “business results”- huh?). I even saw a post yesterday on “Application” Performance Management which suggests to me that we are dangerously close to every business function or process needing its own definition of Performance Management (along with the associated buzzwords and community of followers) for it to be effective. I’m sure there will no doubt be plenty of columns and blogs that will address these, and  the myriad of  new disciplines that emerge  from continued innovation and new technology in the PM arena. No need to waste any more time (other than what I just did!) on that here.

Instead, try to answer the question more directly based on my own experiences. I’ll also use these experiences to reveal the implications of going too far beyond what I believe is that optimal number of KPI’s that should be used to manage the enterprise.

Some “rules of thumb” from my experience…

If I were pressed to answer the question “how many KPI’s are enough?” directly, I would of course  “hedge” a little by opening with the “well, it depends” caveat. But I would still be comfortable laying out some broad rules of thumb that I think reflect the needs of an “average company”.

For example, let’s assume the business has a clear and compelling vision and mission, a handful of established  business goals, and subscribes to the ‘balanced scorecard’ notion of 4 or 5 perspectives (categories or stakeholder groupings if you will) within which goals, objectives, strategies and KPI’s are ultimately managed (i.e. Customer, Financial. Employee, Operating, etc.). See levels 1-3 of the below graphic.

Within that scenario, I would say that the business should have a few key objectives within each of those groupings (left side of Level 4 boxes below). For example, in the Employee Area, one objective might be to “Maintain a Safe Workplace”, while another may be to have “Motivated and Engaged Employees”. My experience is that there are usually 2-5 (max) objectives for each major perspective or grouping. For each objective, you would then have anywhere from 1-3 KPI’s (right side of level 4 boxes below) to measure the success of each.

In my experience, that’s as far as I go with what I call Key Performance Indicators at the corporate level. Do your own math, but my experience is a total of 20 to 30 KPI’s max. Larger companies with multiple business units (especially if highly diversified) may have significantly more than that, while smaller organizations with more limited focus may have less. But again, 20-30 is a good rule of thumb.

A Sample EPM Architecture

That notwithstanding, this is always “context dependent”. For example, if a company decides to replicate this infrastructure in each business unit, the numbers would increase proportionally. But the number of “corporate KPI’s“, those managed at the Enterprise Level would still lie in the 20-30 range

If you’re jumping out of your seat right now, you may be one of those who believes at your core, that KPI’s should be a very small set of things that are “supercritical” to business success (as in 5 or 10, not 20 or 30). I’m not going to take issue there, because I happen to buy into that principle. In fact, I often will extract 5-10 indicators that for the business may be truly “key” or very essential to their success. But for me, I tend to view these as strategic KPI’s or goals that distinguish them from the full suite that I referred to earlier. The full suite (universe)  of KPI’s for the business still falls in the 20-30 range.

Now, assuming many of the KPI’s are “calculated” off of other data or metrics, and assuming that the company would desire to view each by a number of different dimensions (time period, segment, geography, etc.), its easy to see the number of data elements (whether they are viewed as metrics, indicators, or source data) can jump well into the thousands or even tens of thousands.

It’s all part of the “Roadmap”

Ok- maybe I went into the jargon and distinctions more than I wanted to, but I wanted to give you a sense of why I believe the magic number of KPI’s (KEY performance indicators) is where I placed it…in the 20-30 range. And to do that, I guess I did need to define and explain the framework a little. Hopefully, the  chart above gave you a sense of how that architecture fits together.

But even if you buy into the structure I laid out above, there will always be variants. So don’t feel completely bound by the structure or words that I am using here, or the “rule of thumb” benchmark I’ve laid out. The exact number will always be unique to a business environment, but that number should be part of an overall roadmap of how you want to manage your EPM program.

The reason for the type and number of KPI’s you select, should be to create balanced focus and direction, while aligning to your desired end point. Every business needs to have a small but balanced set of management perspectives (i.e. more than one) that they manage within, but not so many that it gets diluted and distracting. Same with the objectives you set, and the KPI’s that support them. There need to be enough to reflect business focus and priority, while stopping at the point at which it creates clutter and confusion. For me, that equates to a KPI number in the 20-30 range.

And while drill downs and analytics will add exponentially to the number of metrics and data points accessed, they still reside in a structure which revolves around those 20-30 KPI’s. There are also some psychological reasons for the why we break our objectives and measures  into convenient little chunks of 3-5 within each perspective, and sometimes smaller chunks within those, but that is a secondary factor that only amplifies the rule of thumb that I laid out.

Consequences of “metrics overload”…

As I said earlier, my experience is that when someone asks me a question around the number of KPI’s they should have, it’s usually because there are in fact too many, and that quantity itself has created a dilution of focus. Or maybe its because the structure within which they reside is losing clarity and alignment. Usually it’s a combination of the two, and one drives the other.

But the underlying cause of all of this is usually lack of preparation on the front end. Perhaps the organization jumped straight into a technology fix for what really is a process and cultural challenge. In these cases, the organization may have procured a tool that only works if you populate it with data. So they rapidly populate their dashboards and scorecards with as many metrics as they can think of, not considering the critical connections, relationships and architectural dynamics at play.

Another reason is lack of ownership for the process. Sometimes it starts well, but without continued governance of the process, each business unit runs with their own vision of what “their” scorecard should look like. Sometimes, a clear and cogent integrated architecture miraculously appears out of the ashes, but more often than not, you end up with several different views of what success looks like. Metrics end up being  misaligned, or worse, conflicting across multiple business units in the same organization.

Whatever the reason, the problems of “runaway metrics” often manifest because there really was no EPM “Program” to begin with (no unifying framework to build upon)–just a set of tools to report and analyze data, and if you’re lucky, a few guidelines for how to use “the tool”. In those environments, its not uncommon for us to find medium to large companies tracking hundreds or even thousands of things that middle and upper management refer to as KPI’s, without any real semblance of an EPM program or platform in place. And no matter what definition or distinction you want to use, that’s way too many.

And it only gets worse…

Complicating all of this, is day to day realities of management reassignments, the natural comings and goings associated with staff turnover, and sometimes major changes in leadership that can (and often should) initiate a change in course, along with the emergence of new navigational beacons and waypoints (i.e. new KPI’s). But rather than changing the structure of their EPM platform, and replacing one set of waypoints with the new ones, they simply layer the new set on top of the old.

At some point, the companies become what I call  “hoarders of metrics”, and before long, an otherwise harmless but impotent process, begins to look like utter chaos.

While the word “layers” often has a negative connotation, they are often useful in establishing an architecture for a solution, and by design, can actually create strength within the system. The architecture I describe above, and the type of “line of sight” thinking I described in yesterday’s post are examples of how this can be used to strengthen your EPM program.

But when the layering is done without a deliberate structure and blueprint, these layers (new metrics on top of old, bad metrics on top of good) can, and often does, cause the system to collapse under its own weight.

Getting it back under control

Here are a few guidelines and “healthy” practices for getting your measurement framework leaner, better aligned, and back into focus:

• Know where you are today-Without getting caught up in all the lingo, do a simple inventory of your process vis a vis the framework I laid out above. Ask yourself, how many objectives you have within each key domain. How many top level metrics are there that support these objectives (things that senior management routinely uses to monitor and talk about)? Are there metrics we track that are redundant and perhaps don’t support any of our objectives? Are some objectives missing the metrics needed to track their achievement? What does that picture look like for you in terms of structure and number of KPI’s, and is some “pruning” warranted?
• Commit to an Architecture – We all acknowledge the advice “measure twice, cut once”. That’s even more important here, as the organization can only withstand a failure or two in trying to get a performance management system in place without experiencing major “cultural fallout”. Continuing to “experiment” with measurement and KPI’s without having an architecture and blueprint to guide that process (even if its a crude one) is setting the stage for many of the above challenges and breakdowns.
• Understand the role of the KPI (versus other parts of the structure)- I’m not talking about getting hell bent on semantics, but I do think there is some value in teaching the organization the difference between a KEY performance indicator, and the myriad of other data and statistical fodder that may be used within the overall system. Use the rule of thumb I laid out earlier to guide and test whether this distinction is sinking in.
• Build down, Not around– If I had to pick a direction to build your EPM architecture, I would say start at the top and work down. Once you are at the KPI level, you should be able to start allocating out accountabilities for their achievement, and then, if your culture supports it, those accountabilities can be disbursed in a measurable way to your staff via individual and work team metrics, from which appraisals and reward structures can then be linked. I differentiate this from the notion of building “around”, in which I mean taking the concept of EPM (a prototype) and repeatedly testing it out in new areas and business functions, without any clear roadmap or enterprise structure in place. While that can certainly kick start measurement activity and get things moving, it can also propagate some bad thinking if the underlying architecture and core practices at the enterprise level are not in place.
• Establish a vetting process for new metrics– Its important to recognize why people develop metrics in the first place. We’d like to think that its all from a noble ambition to help the company improve, out of a pure hunger for data. But it goes way beyond that. People develop metrics for everything from defending their turf to getting their point heard. Since data is now the currency through which corporate truth in now established (which is a good thing), so don’t be surprised when the number of metrics begins expanding exponentially. At that point though, you want to make sure your core system does not get infected with junk; so to prevent that, make sure you have a process or checklist to vet any addition of new metrics into the corporate framework.
• Set aside time for “pruning”-Every strategic planning process, should have a step in which KPI’s and the underlying performance architecture for the business is reviewed. Measures that are no longer relevant, or no longer adding value should be dropped. Unclear linkages upward and downward should be evaluated and strengthened. New business objectives should come into play along with their companion KPI’s, but more often than not, they should end up replacing a measure that may have gone away or diminished in importance.
• Don’t be afraid to “cycle down” a KPI – Sometimes, pruning won’t involve eliminating a measure entirely. For example, if your ambition was to imporve or optimize a measure, and you’ve now achieved the optimal point, it may be time to simply go into maintenance mode and start reporting on a lower frequency (weekly to monthly, monthly to annually, and so forth). Think about how many of your measures don’t change one hill of beans throughout the year, yet they continue to take up valuable “real estate” on your dashboards and the often scarce mind-space of your executive team.

Anyone who does yard work or gardening knows that “pruning”, while it involves cutting back, is really designed to produce a healthier and more vibrant plant, shrub or tree. And rather than producing growth outward (taller and wider), it instead encourages the growth the be “fuller” and often “healthier” in future growing seasons. Next year’s growth fills in those “holes” that may have been unsightly, and encourages a more deliberate  growing and robust pattern. Hence, your long term plans for the garden and landscape start manifesting and coming to life the way they were initially envisioned.

That same kind of annual pruning and renewal process can be just as effective in establishing a healthy growing pattern for your EPM initiative, a pattern that can otherwise get interrupted by the confusion, distraction and conflict caused by an unwieldy and burdensome performance measurement process. And, as with everything from gardening to weight gain to maintenance of our autos, its always easier to manage it along the way rather than waiting until there is a problem.

And with that, I’ll wish you all happy “KPI pruning”!!

The one thing that most everyday drivers fear is that infamous “check engine light”. Unless its during the first few seconds of startup (the point at which every indicator on the dashboard lights up for a few seconds), a “check engine” alert is one of the few that signals that indicate big problems are imminent unless something changes fast…as in, stop the car soon and diagnose, or run the risk of being abandoned on the highway with a very costly repair. If you are someone who doesn’t take your vehicle’s indicator lights seriously, trust me (from experience), this is not one you want to ignore.

Dashboards, Indicators, and Alerts…

There are many indicators around us that alert us to changes in status of a process, and deviation from what may be considered to be a “normal operating condition”. And the place where most of these indicators are visible is on our dashboards. Whether it’s the dashboard of your vehicle, the cockpit display on an aircraft, the bridge on a ship, or a control room in a power plant; it’s the one central place where status is monitored and response strategies are determined, most often by the operator of the asset.

Of course, these deviations from the norm that show up on our  “dashboards” occur in varying degrees, and can signal very different things. A “service soon” indicator light your car dashboard is more “suggestive”, and usually means its time for an oil change or tune up. But you’ve generally got some time before it becomes a bigger issue. A “low fuel” indicator on the other hand, is a bit more significant, and usually means you’ve got a finite quantity of miles left before you are what we might call “SOL” (although I’ve tested this threshold on occasion and can attest to the fact that there is some (albeit small) “cushion” past 0 to rely upon). And then there is the “check engine” light that most often means PULL OVER ASAP ( as soon as safe and practical -but SOON).

I’m not sure about you, but I view the “check engine” light as analogous to to an “airspeed alert” that a pilot might get right before a stall condition, or a traffic alert he gets when another aircraft is within the allowed separation tolerance. You might not yet be “past the point of no return”, but you’re pretty darn close.

Dashboards and Scorecards: Is there a difference?

In the Performance Management discipline, we often hear people refer to the terms “Dashboard” and “Scorecard” rather indiscriminately, with little if any conscious distinction as to what they each connote. I’ve often avoided getting too “wound up” about this, because getting caught up in corporate “buzz phrases” and semantics can cause us to miss the bigger issues at play. But after reflecting on this a bit, I think the differences here are in fact worthy of some discussion. Not because the words themselves are super important, but because it is critical that both components (whatever you call them) need to be part of your EPM solution.

So here are is my take on the critical distinctions between the two:

• Purpose: Dashboards are about helping you navigate the journey. Scorecards are about how successful the journey was.
• Type of indicators included: Scorecards generally contain outcome results, Dashboards are usually comprised of leading or predictive indicators
• Timeframe: Scorecards are periodic and longer term (weekly, monthly, annual trends) in the review horizon, while Dashboards are shorter term and can even be real time
• Reaction– Scorecards should provoke next steps that involve introspection and analysis (drill downs, mining insights, etc.) where dashboards usually are designed to “signal” or “provoke” immediate actions or course corrections
• Targets– Scorecards usually report against a target, threshold or benchmark as a percentage gap and trend. Dashboards generally report metrics within or against tolerance ranges, outside of which signal a required change

As with anything, these are not hard and fast rules, but they should give you a sense of where I personally see the distinctions.

Sure, there are some grey areas here. Outcomes for some, may only be a part of a journey for others. There are also cases where an an outcome indicator might be so important that it is worth tracking in both places- on the dashboard AND the scorecard. For example, some car dashboards have an indicator that tells you what MPG you are getting out of your fuel consumption in real time. But while MPG would normally be an “outcome” metric (i.e.scorecard material), it may also be useful to some of us in watching the degradation or improvement to MPG as we change driving patterns (rapid “gunning” and braking, versus more constant speeds, for example).

Examples for the Fairway

A conventional golf scorecard

A few days ago someone posted about this same topic, using golf as their main analogy. And while I agreed with most of what he said, some of his examples created pause as I though of my time on the golf course.

We both agreed for example that the “stroke count versus par” was what you would always find on a conventional golf scorecard (hard to argue with that!). However, I would also say that stats like # of greens in regulation, puts per green, club distances, etc. should also be part of your scorecard, although maybe at a level or two down the chain. After all, these are things that need to be analyzed and challenged OFF the course (although I have been known to peak at them from time to time during the round). In fact, it is not uncommon for many golfers to track these very stats on their scorecard right underneath or beside their actual stroke count.

But if you put all of that on the scorecard, then what does the dashboard look like? We’ll if we think about it in terms of the 5 criteria I provide above, it would likely  be things like yardages to the hole (what i need in order to make my club selection), wind direction (what I need to shape my shot), # stokes ahead or behind the lead (what I need to manage my strategy), speed of the green (necessary in determining the line and speed of your putt), and the myriad of other factors that are utilized by professional golfers before and during a round. And while many of us may manage the above by “feel”, just take a look at a professional’s yardage book and caddy’s notes and you’ll see what looks strikingly similar to a dashboard (albeit manually illustrated with stray marks and notes). And if you want to spend the money, you can always buy some pretty cool dashboards for your iphone or blackberry.

Some Final Thoughts…

I think some of the confusion between dashboards and scorecards is because metrics are often combined in the same visualization, regardless of whether it is called a scorecard or dashboard. Even automobile dashboards have a place for total accumulated miles. Golf GPS devices enable you to enter score count. As I said before, the distinction is sometimes burry, and often not even that critical.

What’s much more important is whether or not you have the full compliment of indicators you need to manage the business. When most executives ask their teams to develop “a dashboard”, the content of what they are really asking for is unclear. Are they hungry for better tracking of results? Or are they asking for better metrics- those that will enable better decisions and more responsiveness? Or are they simply looking for better analysis of the results?

Unless you understand that, it will be hard to deliver on any of these requests or mandates, regardless of what you end up naming it. In the end, Scorecards and Dashboards are merely visualization tools. What';s more important is that you embed and align the right content into these tools that will enable a clear line of sight between vision, objectives, KPI’s, metrics, and initiatives that tells the complete story and enables those who are in execution roles to be successful.

The bottom line is that you are the designer and architect of the info that is displayed, and so all these distinctions- whether it is between scorecard and dashboard, long term versus short term, leading versus lagging, etc.— are really only important in terms of their usefulness in helping YOU design a system that is relevant and useful within your organization. What you call these things is not near as important as whether the system produces the right outcomes.