Some KPIs are obvious, but important…

When it comes to choosing some KPIs,  careful thought may be needed to align them with strategy (see this article on why you should use KPI Trees to do this). Other measures are in the ‘obvious but important’ camp. In process and production environments it’s fairly safe to say that we care about customer service, cost, speed and quality of production. So it’s not worth sweating too much to build process KPIs from first principles every time we have to come up with a set of measures. Once you have seen a full set of process measures, subsequent sets of process measures are generally variations on a theme.

The six basic steps for creating process measures…

1) Assemble Process Experts and Process Maps in a meeting or workshop

People as pieces of a business puzzle

Having the right people in the room helps in multiple ways –

  • They see many of the underlying process  issues (think ‘improvement opportunities’) on a daily basis
  • The process experts understand the process flow (often this is different from what is documented – 9 times out of 10 the process expert is right)
  • Their buy-in will be crucial when the measures roll-out, so get them involved early to maximise their support for the new process measures

The high-level process maps will be needed in step 4 and provide an essential focus for discussion, these are ‘process KPIs’, after all. These maps should be high-level schematics. If you have machine/process-element speeds (or cycle times) shown by product variant and resource levels on them, even better.

2) Agree on the operational questions you need to answer

Two men and woman connecting with lighted bulb by cord.

Common process questions – Process KPIs vary greatly in the detail, but will generally aim answer the following questions…

Question KPI
How efficiently did we use our materials? Yield (percentage)
How efficiently did we use our equipment? OEE  / OPE  (percentage)
How efficiently did we use our people? Labour Efficiency / OLE (percentage)
How much product did we make/process? Output
What proportion of products were in spec? Quality rate  (% or count)
How many went wrong? Waste (kg or units or percentage)
How many did we have to fix? Rework (% or count)
How many customer complaints did we have? Complaints (number or percentage)
How long did our customers have to wait? – Order lead time (average time)
– Customer demand (volume/hour)
Did we meet orders in full? Order fullfillment (percentage)
How much more could we make? Plant/process capacity or utilisation
How much does it cost to make/process an item? Activity based costing (£/$)
How many of our customers gave up in frustration? Call abandon rate / Order cancellation rate
  • Use this selection matrix to keep the list manageable
  • In some cases, several these questions will be answered by one ‘core’ measures, like OEE, yield and so on.
  • There are often process and product-specific measures for a given industry. These normally fall under ‘yield’ and ‘efficiency’, but are helpful to break out separately. Examples might include ‘giveaway’ in the packaging industry (putting more than required in a packet) or ‘call abandon’ rate in a call centre environment. Both affect yield and are a specific type of (avoidable) loss.

3) Arrange the measures in a hierarchy of a KPI Tree

KPI Tree Snippet

  • This will help later when you need to build the dashboards and reports. It will also help you spot any gaps in your coverage.
  • Remembering to focus on outcomes and results measures (at a tactical level) in groupings such as:
    • ‘Lowest unit production costs compared to our top 3 competitors’.
    • ‘Never lost a customer before a ‘decision in principle’.
    • ‘Best lead time in our sector’.

Try and avoid grouping by broad generic categories with no targets e.g. ‘Quality’, ‘Process’ and ‘Efficiency’. These aren’t results and won’t move you towards you strategic goals in themselves.

4) Map the KPIs to measurement points on each process map

Illustration of businessman with eyes closed touching sensor screen

This step determines where in the process the measurement will take place (e.g. which bottleneck speed dictates the process potential for the OEE calculation)

  • You may need to do some process measurement to do this, measuring cycle times or maximum output rates for example

Caution: With measures like OEE,  getting the right kit identified as the bottleneck is critical. Getting this wrong can throw the accuracy of your efficiency off by an order of magnitude.

5) Agree on the specific definition of each KPI

two businessmen holding a sign protesting with different opinions

Many of your more common measures will be ‘off the shelf’ measures, like OEE and Yield. Even these will require some clarification. In some processes there are unavoidable losses (shrinkage in paper making, for example), there needs to be a discussion about what is, and is not, included. For the record, I tend towards ‘brutal’ measurement – leaving things ‘in’ even if they are seen as ‘part of the process’. This is because some of my most successful manufacturing consulting results came from reducing the ‘part of the process’ losses that could have been hidden if they were excluded from the yield or efficiency measures.

6) Reverse brainstorm your new process KPIs

Machine stealing idea from man's head and converting it in money. People in queue with illuminated bulbs

Use the process experts to answer the question ‘Could we get these measures to go in the ‘right’ direction but still get a bad outcome for the business?’

If there are missing measures identified, skip back to step 2 or 3 and add them in for each of the following steps.

Note: Many of the slightly more ‘exotic’ process KPIs are variants on the answers to these questions. For example – Capacity planning is simply the combination of a forecast with a fine-grain capacity/utilisation answer.

What about financial and safety measures?

Of course there’s a whole raft of cost-revenue and safety focussed questions we could also ask. From experience these are generally the best covered measures in any business, but if you need to build some finance or safety KPIs you can do it in just the same way but using modified questions in Step 2.

The journey of a thousand miles starts with a single step

As with all KPIs, it can be easy to plan them but much harder to implement them. If process KPIs are a fairly new idea for your organisation, start with something fundamental. For those completely new to process measurement, just measuring daily output can deliver a lot of insight and value. For more advance operations, OEE and Yield are early ‘must have’ KPIs. Focus on getting the implementation right, rather than rushing into too many different measures.