Wednesday, 16 November 2016

Getting the most out of KPIs

There are two old adages, "You can't manage what you can't measure" (from Peter Drucker) and "What gets measured gets done" (provenance unknown) which apply as much to strategy as to most other things.

That is why I always advocate that you articulate your strategy all the way through to measurements (and targets).

A simple wat to do this is to translate your strategy and vision into goals, break those goals down into SMART objectives, and create measurable KPIs and targets to describe those objectives.

What sounds simple in theory can be extremely difficult in practice, and I see a lot of companies give up altogether, or build dashboards out of the data that is readily to hand whether or not it describes their strategy.

Type of KPIs

In addition, not all KPIs are created equal.

For example, you get:
  • Quantitative KPIs: represent things you can measure using some unit of measurement. Examples include Sales (measured in units sold) or Sales Revenue (measured in monetary terms).
    • Synthetic KPIs: are a subset of Quantitative KPIs where instead of measuring something, you create a score which may be a composite of a range of measureable phenomena. For example, Google's PageRank is a synthetic score calculated for a web page based on the PageRank of every page linked to it combined with an authority score for the author of the page. Net Promoter Score (or "NPS") is another example calculated using a defined formula determined from customers' answers to a simple question.
  • Qualitative KPIs: represent things you can categorise and rank. Examples include high, medium and low or variants thereof, or initiatives which are recorded as moving through stages (initiation, analysis, design, development, testing, deployment live, etc.)
  • Binary KPIs: represent things that either are or are not. Examples include initiatives which are shown as being on track (or not), or complete (or not).

Information Content

Different KPIs provide different amounts of information content. Quantitative KPIs, especially very granular ones, tend to have the highest information content. Qualitative KPIs have less information content: if a KPI is high, how high is it? If it still high in the next measurement period, is it higher or lower than it was the previous time? Likewise, if a KPIs shows a stage in a process, how close is it to moving to the next stage? Unidirectional binary KPIs contain the least amount of information: for an example, an initiative is not done, right up until the point it is done, after which it can never be not done again - it will ever only yield exactly one bit of information.

For this reason, I always prefer quantitative KPIs over qualitative KPIs, and I avoid binary KPIs, especially unidirectional binaries, as much as is humanely possible. Invariably, I find that qualitative KPIs and binary KPIs can be translated into quantitative KPIs, but that is perhaps a subject for another post.

Objectivity and verifiability

A second consideration in selecting PKIs is how objective and verifiable they are. For example, quantitative KPIs which come directly from a business's independently audited financial statements tend to be quite objective (although not always very strategic).

Subjective measures can be made more objective (and often more quantitative at the same time) by average the opinions of large numbers of people (this is how Net Promoter Scores work, for example).

Qualitative or binary KPIs based on one person's (or a small group's)  opinion, on the other hand, are much more open to bias. I am sure we have all seen initiatives declared as complete by politically motivated sponsors when everyone knows that key outcomes have not been delivered.

Including projects on strategy scorecards

As a category, KPIs for the completion of projects deserve special mention. I am personally not a fan of including these in strategy scorecards because (1) the usually have low information content (as outlined above), (2) quickly become political and subjective (also as outlined above) and (3) you either clutter your scorecard by including every initiative in the organisation, or are forced to make subjective choices about which to include and which to exclude.

Much better I believe, is to focus the strategy scorecard on the strategic outcomes to be achieved by the projects. This has the added benefit of rewarding agile projects which deliver some benefits early, over higher risk 'big bang' projects which backend all of the benefits and which are problematic in their own right. As a side note, if part of the organisation's strategy is to improve its execution / project delivery capability, then I think there are more than enough quantifiable capability metrics that can be aggregated across all projects.

Conclusion

KPIs are powerful enough to make or break a strategy, and they are notoriously difficult to get right. They can be used to minimise the interference from (corporate) politics, and drive execution. Extreme care and diligence should be applied when setting them.