Chick Operating Officer

Seasoned operating executives will generally agree that a well-run business needs to have a set of metrics, above and beyond bare-bones financial results, that are used to guide decision-making and measure progress.  Without metrics we would be flying blind.  Unfortunately, operating metrics can also look like a panacea to the uninitiated, thereby lending themselves to fad-surfing.  When boards, investors, and far-removed executives are at a loss about how to improve business results, demanding that metrics be put in place has a certain simplistic appeal.  A metrics initiative can create the illusion of taking action, it sounds easy enough, and it might actually fix the business.  Plus, metrics can provide a means of avoiding the need to assess people and strategies via more time-consuming and uncomfortably subjective means. What’s not to love? 

Longfellow presumably wasn’t thinking about metrics when he wrote this line, but it is particularly apt: “And when she was good, she was very good indeed, but when she was bad she was horrid”.  Metrics are necessary, useful, and sometimes transformational – but a mismanaged metrics initiative can go horribly wrong and end up harming the enterprise.  In the ‘very good indeed’ case, metrics are woven into the fabric of running the business.  The information is used on a day-to-day basis, in an integrated manner, by line managers who understand what the graphs are telling them.  They use the data to tweak performance on an ongoing basis, and the same information can help other stakeholders assess progress.  This is in stark contrast to a ‘horrid’ metrics scenario, in which the organization is whipsawed from one extreme to another as it attempts to optimize various parameters.  And if the culture leans toward blame and punitive behaviors, metrics become the new stick with which to beat people.    

So what can be done to maximize the chances of success when designing and implementing a metrics program?  The choice of specific metrics is highly situational – there is no magic formula on that front.  However, there are some broad practical principles that can help guide the process.  Here are a dozen points to consider. 

1. Prioritize and pare down.  The word ‘dashboard’ has been appropriated to describe displays of metrics information, and it’s a nice analogy:  a few key pieces of information, graphically displayed, to aid in navigation.  Bizarrely, however, many ‘metrics dashboards’ seem to consist of multiple pages jam-packed with numbers printed in a 4-point font.  Modern ERP systems make it easy to generate reports, but they also make it easy to add extraneous data – and this information overload can obscure the most meaningful numbers.  The key to an effective metrics initiative is to focus on a limited number of judiciously selected measures, with a hierarchy of supporting detail available for deeper dives.  You can always add things later, but keep it lean to start with. 
2. Align data reporting with data velocity.  There is also a time dimension to information overload.  Consider the following example.  In a product business that relies on non-recurring revenue, it is probably important to generate a flash report every morning to provide orders and sales information.  But what is gross margin doing in there?  Daily GM fluctuations probably aren’t meaningful, even if they are measureable.  Carefully considering the frequency with which things are reported is a critical aspect of achieving focus and de-cluttering reports.
3. Paint a holistic business picture.  Financial metrics from both the income statement and the balance sheet, as well as classic operations metrics (quality, throughput, labor utilization, etc.), are obvious components of a metrics program.  In order to maintain balance, consider also including quantifiable views of human factors, such as results of employee satisfaction and/or customer satisfaction surveys.  Things like percentage of revenue from new products, although always a bit imprecise, can serve as a useful leading business indicator.  And external data, such as market evolution as measured by new technology adoption, might also be included if it represents an important driver of the business. 
4. Consistency is key.   The basic idea of metrics is to observe and manage performance over time.  While the need to focus on trends is readily grasped, the importance of looking at exactly the same thing over a period of time can sometimes get lost.   It can be tempting to present slightly different information during each review cycle, perhaps because a more relevant calculation has been genuinely identified, or perhaps in an instinctive attempt to present good results.  (Something is always going up and to the right.)  Review and make adjustments periodically, but do it in full view and not too often. 
5. Use external benchmarking thoughtfully.  If you can be fairly sure of an apples-to-apples comparison, benchmarking to your direct competitors or industry averages can be an important part of a good metrics program.  For example, I’m a big fan of knowing exactly where you stand with respect to expense ratios for various functions.  There may be a good reason that you are spending a lot more (as a percentage of revenue) on engineering than your main competitor, but you should be very aware that this is the case.  A counterpoint is the once-popular ‘revenue per employee’, now rendered fairly meaningless by widespread outsourcing.  Choose carefully and try to avoid fads.
6. Proceed cautiously with formal metrics methodologies (e.g. Balanced Scorecard).   In a highly structured environment with dedicated resources, these academic approaches can presumably be effective.  However, it has been my experience that they don’t work well in smaller companies.  I have seen metrics initiatives simply collapse under the load of excessive complexity and administrative overhead.  Reading up on BSC to glean some of the underlying principles isn’t a bad idea, but in general I recommend avoiding gimmicks.  Likewise don’t put a huge amount of energy into selecting and implementing a visual dashboard package.  Devote your resources to choosing a good set of metrics, and then proceed with a very basic regimen of regular reporting and review.   
7. Also use caution with financial metrics based on calculations (ROIC, EVA).  These measures of financial performance, which typically strive to incorporate both profit and capital deployment considerations, can be very useful in measuring a business.  However, unless people can really understand the formula and develop an intuitive feel for the factors that impact it, a focus on one of these measures won’t produce tangible results.  If you decide to go this route, limit yourself to one key measure – and be sure to explain it thoroughly and often throughout the organization. 
8. Involve the whole organization in the design phase.  A purely top-down or finance-driven exercise is unlikely to be successful, simply because an intimate knowledge of the business is needed for good metrics design.  While executive sponsorship and coordination are of course critical, a bottom-up metrics design is more likely to focus on factors that can really have an impact.  It’s also an important means of establishing shared goals and achieving buy-in. 
9. Foster a culture that supports continuous improvement.   Within a troubled business culture, it is probably possible to use metrics to achieve some marginal operational gains. The net impact is likely to be negative, however, since a dysfunctional culture will become even more so with the implementation of metrics.  More weaponry for assigning blame!  And don’t kid yourself that it’s possible to design a metrics system that can’t be gamed.  In an environment with serious behavioral problems, my recommendation is to work on the cultural underpinnings before implementing a metrics program.   Utopia isn’t necessary, but people need to believe that the tools are intended for collective business improvement and not for punishment.  This requires some modicum of trust, open communication, and management skill.
10. Develop an internal metrics communication plan.   In order to make incremental improvements, metrics need to be looked at regularly and understood by the people who can affect the inputs to the equation.  Old-school manufacturing leaders understood this well, and they addressed the issue by hanging big quality graphs on factory walls.  It’s a bit harder to achieve in today’s fragmented companies, so be sure to carefully consider how, when, and where the information will be disseminated into the organization.   It’s going to take more than an occasional email attachment.              
11. Remember the observer effect.  When selecting metrics, keep in mind that the act of collecting data may affect behavior.  For example, my experience suggests that an attempt to compile won/lost data from sales is likely to reduce the visibility of sales opportunities.   Even in a healthy culture, no one wants to be a loser—so why report a pending deal if it has the potential to result in a black mark?  Worse yet, the sales team might actually shy away from pursuing long-shot opportunities at all.  In the final analysis, it’s probably more useful for the business to maintain unfettered information flow about potential deals than to have won/lost data.  (Also note that many metrics have no such backfire effect.  For example, sales linearity can often be improved by a metrics initiative.) 
12. Reject metrics that don’t help you run the business better.  This one is self-explanatory, and perhaps the most important point of all.