Portfolio Management with Starbucks

I've been trying to design a game which will teach some agile & lean principles around running a portfolio of work.. And to be honest I've been struggling!! So many learning points to get across whilst making it fun.

Anyway - I've been doing some work today with some teams that are using SAFe..  Upon arriving back at Euston I find my train is delayed.... Yet again! So I decide to spend 42 minutes or more in Starbucks.....

Whilst watching the queue in Starbucks I observed that just about every behaviour to explain portfolio management can be learn't from watching the queue in Starbucks!

For those familiar with Little's law - The time between joining the queue and being served (delivery) is relational to the size of the queue (Lead time). Okay.. Nothing surprising there!

However the speed the queue moves at is variable..... But why.. Well if everybody in the queue orders a single cup of tea (And that's all you should order, A nice tea) the queue moves at a predictable speed... However if most people are ordering tea and then somebody comes along and orders 6 Chai latte's for everybody in their team back at work all of a sudden the queue stops moving! Variability has been added to what was a predictable system.

And then... Somebody joins the back of the queue and spots a friend near the front.. So what do they do? They add their order to their friends order which caused even more uncertainty and irregularity to the flow!! Okay so I made that last bit up, That never happened! But how many times with projects do you see that? PM's buddy up with a mate and piggy back part of their projects onto somebody else's?

Then when Starbucks see that the queue is getting long... What do they do? They send somebody out to take everyones order! This really annoys me has It hasn't nothing to do with speeding up the queue.. Taking the order isn't where the constraint in flow occurs! The constraint it usually the coffee machine(s) which are optimised for 2 drinks at a time - Taking people's order is merely to keep them in the queue and to stop them wondering off somewhere else... Sounds familiar?

I did some consulting work on a eCommerce product a while back and one of the major objectives was to reduce lead time.... So.... How did I achieve this? Well it's really easy.. I just shortened the queue!  Short queue equals short lead times!

It was a tough policy (not one that I expect Starbucks to implement)... The queue couldn't be more than 3 months of work! (Well not totally true... but almost) once full nothing else could be added! The result meant that if you got your idea into the queue you knew it would be implemented within a maximum of six months... Usually less than three.

Every three months projects would need to be reviewed (including any in the system)... The benefit (business value) re-examined and fought out with other projects... The projects that were SMALL with the most business value won and were allowed to join the queue! If your project wasn't small and you really needed it done.. Well you had to find a away to make it small enough and to still deliver value!

I still haven't designed my portfolio game.. Perhaps instead I should invite them out for coffee instead.


For more info on Little's law.... 



For those wondering...... Starbucks has not sponsored this post!