Nelson Repenning, Associate Professor of System Dynamics and Organization Studies, Director, MIT Executive MBA, MIT Sloan School of Management
Nelson started by asking the room – who is less busy this year than 10 years ago. No-one put their hands up!
Despite everyone talking about automation, no-one is less busy! Particularly for knowledge workers…
The costs of overload are well documented. Overloaded organisations are less innovative, develop fewer products, have higher turnover, and are prone to accidents and warranty problems. Despite this, it is ubiquitous.
Broad Institute, MIT
They are looking to unpack the data encoded in cancer genomes, with big data as the output. They have made huge gains in productivity in two different ways. First the problem.
Nelson’s student Sheila became director of the organisation. When she joined the organisation was facing several problems – 1) decreasing demand due to competitors and 2) increased, unpredictable turnaround times with 3) decreased machine utilisation of around <50% (the sequencing machine is the most expensive piece of equipment with a relevance time of only 12-18 months before requiring update).
Push vs Pull systems
They used a push production system – a sample would appear, get prepped and immediately go to the lab ready for processing. It would continue through these steps until making it to the sequencing machine. The problems with push are 1) it is slow – the more inventory in the system, the longer from start to finish 2) there is local reprioritisation where different people prioritise differently leading to variable cycle times 3) there is expediting where people work around the process – eg people calling to chase their samples, or friends’ samples getting to the front.
Expediting forms a pathological downward spiral where 1) the rest of the samples are late, leading to further chasing and 2) wasted time searching out the samples.
How to fix a push system? Move to the pull system. 1) Restrict the work in progress inventory – in Broad they used coloured boxes in the refrigerator to sort work. 2) Only prioritise once – you can reorder the work at the very start, but once it is in the process, you can’t touch it! This lead to 1) faster cycles, with 50% reduction in time and 2) more predictability 3) the utilisation was over 90% – as no more chasing of samples.
This has lead to business growth as measured by increased revenues.
So how can you “pull” knowledge work?
The majority of knowledge systems work in “push mode” – eg email inboxes. Nelson has used a technique called “Visual Management” to give knowledge a physical manifestation. Initially, they used post-its and markers.
The way to think about this can be summed up in “mental scaffolding” – from an evolutionary perspective, knowledge work is very new. We have dealt with physical “stuff” up until very recently. Therefore it is not surprising that we may not be as well adapted to dealing with these intangible assets. There are huge amounts of research being done on this topic in psychology.
Mental scaffolding – our understanding of abstract concepts is built on the scaffolding provided by physical experience (see Ackerman, Huand and Bargh).
Ackerman did interviews where interviewers held a cup of warm coffee and rated job candidates as “warmer”; when holding a heavier clipboard rate candidates were rated as more “serious”.
The takeaway is that – if you buy this argument – it’s not surprising that we came to treat knowledge work in the same way as physical work because of this scaffolding. In factory work, you can see the work stop because the line stops – how can we take that rapid feedback that we get so naturally from the physical world and translate it to knowledge work.
Visual management is “pull” for knowledge work.
When the board looks like chaos – the organisation is chaos!
The magic is in the conversations in front of the board, rather than the board or content itself.