Lean Implementation – When to Pull

[tweetmeme source="leanisgood" service="ow.ly"]

Kevin Meyer at the Evolving Excellence blog does a nice series called 5 Questions in which he asks various people associated with lean five questions (good name for this series). I find myself implementing either lean or continuous improvement at my third organization and something that Michael Balle talked about a few months ago when he was answering Kevin’s 5 questions got me thinking (here is the post with Michael’s 5 answers to Kevin’s 5 questions) about when you should begin perfecting your value stream (create stability, eliminate waste, etc) and when you should ‘connect’ your value stream.

Although I recommend that you read the post in its entirety I would like to focus on part of Michael’s response to question # 4 – In your opinion what is the most misunderstood or unrecognized aspect of lean?  Quoting from Michael’s response (emphasis mine):

The second aspect of lean on which we tend to run in trouble is insisting on pulling ASAP, rather than “when we’re grown up”. There are all sorts of so-called “lean” arguments for not pulling: not enough stability, unclear flows, too large batches, too complicated and so on. I actually used to believe that and six or seven years ago disagreed with Freddy on the “pull as fast as you can” tack – I believed management control and stability took priority. I have now to accept (and confess) that the sites where they took this stance never ended up pulling and their improvement activities keep waxing and waning according to management energy. On the other hand, if we commit to pulling right from the start, we can immediately start levelling, clarifying value streams and reducing batch sizes – which starts “the river and the rocks”; The pull system gives an architecture to kaizen, which sustains further improvements. Every program I’ve seen that does kaizen without pulling ends up exhausting itself.

The now famous boat graphic (click to enlarge)

If you think about the ubiquitous boat on the water graphic that most of us have seen in one incarnation or another a bunch of times (the one where the boat is our process, the water represents our inventory, and the rocks and shoals represent the various instabilities in your current process), you can visualize the advantages to Michael’s (and my) approach to implementing lean – PULL AS SOON AS YOU DARE.

Establishing pull (and flowing where you can getting flowing quickly) early in your implementation will do a couple things for you:

1. It allows you to focus on the really important couple of things.  Before pull and flow your inventories probably jerk up and down very quickly and very frequently.  On days when your problems (instabilities) are underwater then everything is good.  On days when your inventories get very low then you are confronted with a bunch of problems all at once.  The stability in inventories created by pull will show you what your vital few problems are.  In our example above long set up times are our first problem and maybe we bump the bottom of our boat on poor housekeeping from time to time but the 800 pound gorilla will be changeovers and we get get “all hands” working on that.
2. It will find the level of inventories that you actually need today in order to service the customer.  Many times when implementing pull people have to increase inventory levels (remember the original intent of pull was to assure that there was enough inventory to service the customer).  People will often artificially drive inventories low in their MRP systems to meet arbitrary inventory reduction goals.  Pull (and specifically kanban) will put a system in place which will allow you to reduce inventories over time.
3. It brings inherent advantages over push scheduling.  These include but are not limited to (I invite you to add more in comments below): avoid over-production by replenishing actual consumption instead of building to a forecasted model of demand (reality is more accurate than a guess no matter how educated except in the special case of luck), pull is a self-adjusting system for variations in demand and lead time within constraints imposed by your selected ‘safety stock’, pull provides buffers for error propagation at selected points in the value stream (supermarkets) which mute (if not moot) the bullwhip effect from gain amplitude as it tsunamis its way through you value stream, and it is generally simpler than most forms of push which I suggest is an advantage.

A former boss of mine used to talk about getting a ‘connected value stream’ before obsessing with perfecting.  His argument was the same as Michael’s.  Bryan challenged me to help him with end to end pull (integrating vendors and customers) in a very narrow value stream when he was a Value Stream Champion and I was his sensei in a previous life.  Womack and Jones dub the five lean principles: Identify Value, Map the Value Stream, Create Flow, Establish Pull, and Seek Perfection.  They consistently present them in that order.  I can’t say that they intend that to be a prescriptive schedule for lean implementation but they have maintained the integrity of that order throughout the years.  I think that is a pretty good order though because pull comes before perfect.

I say “be bold and pull early!”  What say you?

Bruce

Advertisement