The contrasts in the June issue were fun, showing that lean service is still an area of great debate. Better that way.

Tracing the flows through Lethbridge’s matrix of 16 elements could lead to hours of discussion. Good on you for getting us going, Sarah. But ‘standards’ or standard work; variation or variety? Also very useful to be reminded later by Squires of the value analysis view that value is function divided by cost. So, in services, if we think of waste being something that the customer would not pay for (see Gillet’s article), it depends on cost. Perhaps then a case of the Kano model of basics, performance factors, and delighters. This leads onto…

Standard work: a contrast between the ‘standardisation spectrum’ of Lethbridge and Bill Bellows’ warnings about firefighting on the one hand, and the (apparently) more rigid views of Gillet and Locher on the other. So, is there a compromise?

Let us first reflect on possible confusion between standard work and work standards. Standard work should be more about the CURRENT best way to do a job (and hence evolves and is developed by people at work at the gemba), whereas a work standard defines what is required as an output (and hence tends to be more static, and is defined by customer need through managers or engineers).

Then, I think, it is helpful to look at utilisation, which we know has a major impact on delay and throughput particularly where it is too high. This is muri or overload, and the negative consequences begin well below 100% utilisation. For people as well as machine. At high utilisation the effects are exponential.

Utilisation is load divided by capacity. Load, in turn, is value demand plus failure demand. So any activity that leads to failure demand increases utilisation and must be avoided, particularly at high utilisation levels. Then there is capacity. Capacity is base capacity minus waste. So if waste can be reduced, capacity is freed up, utilisation drops, and delays reduced.

Even Starbucks (see Roberto’s article) with attention to standard work, still has complaints. That is why they teach their LATTE recovery procedure – Listen, Apologise, Take action, Thank, Explain.

The other factor in Kingman’s equation is variation. There are two types: arrival variation and process variation. Much of the discussion in the articles was about process variation. Here, standard work (and six sigma and 5S) could play a role provided that failure demand is not increased. A very big proviso! Better to absorb than try to reduce? Arrival variation is another story, but equally important. Here the article on Toyota service was interesting – influencing arrival variation. For each type of variation, it is relative variation that is important, not absolute. This means that variation in a long process is unimportant, but variation in a short cycle process is very important. Would that make the case for combining front and back office, rather than separating them in the interests of efficiency? Particularly if failure demand results.