Planning Mark 2: Doing a Dynamic Projection based on current info
After the not so smooth-running approach with her Gantt chart, Betty The Baker does not even try anymore to give the illusion that she can plan half a year in advance.
Instead, Betty remembers that she can answer the question “how long does it take”, and that the order volume is actually quite stable (if you want to show off, you could call this the cadence). Also, Betty decided she wants to allow her customers to change their minds until the last responsible moment (but not later!)…because this flexibility is her USP, which she’s very proud of.
So what’s the solution? During her next weekly meeting, Betty cuts slices through her backlog, one for each iteration. Since the backlog was already sorted by relative priotization, the most important tasks are on top, and will receive an early slot. Lower-ranked tasks will also receive time and attention…just a bit later.
This way, every Monday, the bakery can now tell all the customers which iteration they are currently in, and where they stand in line. Remember our two examples from earlier – we’ve already seen how “place in line” relates to your individual waiting time.
It adapts!
Contrary to the static Gantt chart, this new planning approach is self-regulating, and can incorporate new information every week. We call this adaptation, a principle stolen from biology.
For example, if Betty’s mixer breaks down, she’d need to mix by hand (or make her shop assistant do it), making her slower. Therefore, for all following iterations, her capacity is reduced.
But since our planning has the superpower of adaptation and is based on capacity, all the answers to the “How long?” immediately factor in the new capacity. If you’re curious and want to read more on this, check out Queueing Theory (you know, the baking speed can actually be computed) and Little’s Law, which is about the relation of queue length and work in progress.
I want you to do me a favor though
We did it! Planning! Now Betty the Baker is able to answer every customer’s question when he or she will receive their orders. Buuuut since our baker has a lot of buddies and a large family….what happens if they all come into the shop and are served first?
Let’s say this cheating the line happens on a Wednesday, what impact does this have our adaptive planning? We assign “slots” to the customer orders every Monday, so after cheating, we’d squeeze more work into our week then capacity allows. Some jobs will remain undone, even though we promised to ship it! How would you feel if you were standing in line?
We’d be back to the dark side of being industrious, which is multitasking. We start working on quite a lot of orders, but never get around to actually finish them. So, most orders just sit there, waiting, and all info we gave our customers regarding the “when is it ready?” eventually comes to naught.
So at that point, there’s one final but super important bit missing in our planning approach. After our Monday meeting is over and our iteration starts, we must not add or remove orders.
So, to summarize, here’s a quick overview and a mapping of our bakery concepts to actual Scrum.
| Concept | Example | Scrum |
|---|---|---|
| Base Unit | Bread | Story |
| Iteration | Week | Sprint |
| Estimation Unit | Person Hours (Fixed at 1/Piece) | Story Points |
| Capacity | Capacity “Bread per Week” | Velocity “Story Points per Sprint” |
| Daily Check-in Meeting | Daily Meeting | Daily Meeting |
| Planning and Improvement Meeting | Weekly Meeting | Sprint Planning, Sprint Review, Retrospective |
| Effort Estimation by | Experience | Planning Poker |
| Role “Baker” | Baker | Development Team |
| Role “Business Owner” | Baker | Scrum Master |
| Role “Customer” | Customer | Product Owner |
Famous last words! For those whose brains are idle — How does inventory size relate to the stability of demand? What would be the ideal inventory size? If you’re curious about that, check the sequel post “The Bread Game” here.