Multi-tasking fails The cost is invisible The choice is unconscious Clarke Ching, Agile Expert, GE Software Solutions Group with Andrew Ireland, Lean Leader, European EED Technical Centres

Multi-tasking is Evil. Hi, it’s Clarke here again. In the next 5 minutes or so I want to show to you just why multi-tasking is so bad for business. First, I’ll introduce you to 3 very simple projects then I’ll execute them under 3 different scenarios: Ideal scenario – each project has a dedicated team so they don’t share resources. More realistic/worse – reduced resources forced to share between projects. I’ll choose to execute the projects without multitasking. Default/Considerably worse – reduced resources multitasking between projects. Many organisations do this without choosing to – it’s the default. And then I’ll compare the results in terms of lost profits. It’ll only take 5 minutes but by the end you’ll see that multi-tasking is (a) a choice (b) a very expensive choice, and (c) a choice we often make without knowing

Assume we have three projects I used to demonstrate how context switching robs many of us of capacity by adding on, say, 20% extra effort to each task. Today we are going to ignore the cost of context switching. Why? Because, counter-intuitively, it’s only a small part of the true problem. To see why let’s pretend that these three projects are real and that each letter/digit/numeral is a 2 week task. So – all things being equal – we have 3 projects requiring 20 weeks effort each, like this: A B C D E F G H I J 1 2 3 4 5 6 7 8 9 10 i ii iii iv v vi vii viii ix x

But, now, to make this more realistic we need to complicate things a little by pointing out that we have three types of specialist resource: Blue resources specialise in the earliest 3 tasks – A,B,C, 1,2,3, i,ii,iii Yellow resources specialise in the middle 4 tasks and, Green resources do the last 3 tasks. So, our projects look like this: Let’s also assume that each project has it’s own project manager. There is one more, very important, thing we need to discuss to make this real: Why are we doing these projects?

Let’s say that these three projects fulfil real business purposes with real financial benefits: These $$$$$ are the BENEFIT of finishing the projects. You can also think of these numbers - $10K, $8K, and $5K - as the COST of DELAYING these projects by a week. Project 1 A new product which earns $10,000 per week, once live. Project 2 Once delivered, saves $8,000 per week on postage. Project 3 Fixes a known defect in one of our “live” products which currently causes support costs of $5,000 per week (& decreased customer satisfaction). We’ll save $5,000 per week once delivered.

Ideally, we’d have 3 of each type of resource - blue, yellow, green - so that each project could proceed independently of the others. That way they’d finish as soon as possible and make as much money as possible. Let’s see what looks like on the next page. Ideal Scenario

1. Ideal scenario, no shared resources P1 P2 P3 * Each task takes 2 weeks to complete x3 x3 x3 Resources: Each project proceeds independently . They’ll finish at the same time. Cash Flows at the same time.

Sadly, in the real world we have to share resources.  Say, we only have 1 of each type of resource. We still have 3 project managers. And they each want their projects to start immediately and keep moving. But … we also have a commercial manager who prioritizes the projects to maximise cash flow. Let’s call this person a FLOW-MASTER. Flow-masters hate multi-tasking Shared Resource, No Multi-tasking

And then, after handing off to Yellow, who starts work on P1, Blue moves to P2, the next most important project. This is single-tasking! 2. Shared-resources, Single-tasking A x1 x1 x1 B Resources: P1 C D E F G H I J P2 P3 1 2 3 4 5 6 7 8 9 10 i ii iii iv v vi vii viii ix x 20 weeks 28 weeks 36 weeks $10K/week $8K/week $5K/week The flow-master wants the cash from project 1 to flow first. So Blue sticks with P1 until complete. Notice the cash flow:

Let’s take away our FlowMaster and let the project managers prioritize between themselves. So we still have 1 of each type of resource. We still have 3 project managers. And they each want their projects to start immediately and keep moving. How do they prioritize? In the name of “fairness” they agree to share resources … asking the staff to “multi-task”. Shared Resource, Multi-tasking

3. Real World Bad – Multi-tasking A x1 x1 x1 B Resources: P1 C D E F G H I J P2 P3 1 2 3 4 5 6 7 8 9 10 i ii iii iv v vi vii viii ix x 48 weeks 50 weeks 52 weeks $10K/week $8K/week $5K/week Notice the cash flow: How is this going to work? In order to keep all PMs and their customers happy, Blue will work on P1 for a bit, then P2 for a bit, then P3 for a bit, then P1 for a bit, then … Notice how all three projects are making progress, albeit slowly? They are multi-tasking.

Shared resources, Single/Multi – tasking comparison On the following pages we’ll compare the effects of the last two scenarios. You’ll see that I’ve hidden the tasks – as a business we don’t really care about them. We’re really only interested in the end-dates. Oh … and the cash flow.

Compare Project 1 A B P1 C D E F G H I J A B P1 C D E F G H I J 48 weeks Multitasking No Multitasking Notice that P1 finished (an incredible) 28 weeks sooner. Cash starts flowing 28 weeks sooner. At $10K per week the cost of multi-tasking – the lost profit - was $280,000 20 weeks Let’s look at Project 1, which has a cost of delay of $10,000 per week.

Compare Project 2 P2 A B C D E F G H I J P2 A B C D E F G H I J 50 weeks Multitasking No Multitasking P2 started later … but finished 22 weeks sooner. Cash starts flowing 22 weeks sooner. At $8K per week the cost of multi-tasking – the lost profit - was $176,000 28 weeks Let’s look at Project 2, which has a cost of delay of $8,000 per week.

Compare Project 3 P3 A B C D E F G H I J P3 A B C D E F G H I J 52 weeks Multitasking No Multitasking 36 weeks P3 started much, much later … but still finished 16 weeks sooner. [Go back and check … I just did.] At $5K per week that’s $80,000 lost forever due to multi-tasking. Let’s look at Project 3, which has a cost of delay of $5,000 per week.

Compare Project end Dates Notice All 3 projects finish much sooner. Yes, all of them. A B P1 C D E F G H I J P2 P3 A B C D E F G H I J A B C D E F G H I J 20 weeks 28 weeks 36 weeks A B P1 C D E F G H I J P2 P3 A B C D E F G H I J A B C D E F G H I J 48 weeks 50 weeks 52 weeks Multitasking No Multitasking

Compare Throughput Without multitasking we can complete 2 more projects of similar size A B P1 C D E F G H I J P2 P3 A B C D E F G H I J A B C D E F G H I J 48 weeks 50 weeks 52 weeks Multitasking No Multitasking

Compare profits And, the multi-tasking scenario made $536,000 less profit than the non-multitasking scenario. $536,000 lost forever A B P1 C D E F G H I J P2 P3 A B C D E F G H I J A B C D E F G H I J 20 weeks 28 weeks 36 weeks A B P1 C D E F G H I J P2 P3 A B C D E F G H I J A B C D E F G H I J 48 weeks 50 weeks 52 weeks Multitasking No Multitasking

Nearly there … I’m going to finish shortly. But did you see how multi-tasking – in this simplistic and contrived example – cost us a lot of money? In my experience, most organisations multi-task. And most do so without realizing that it is a choice and that it has a very high cost. The cost is invisible. The choice is unconscious. It’s very similar to how non-lean factories work in big-batches, not realizing there is a simpler, better, more profitable way. The first step to fixing a problem … is realising you have a problem. Does multi-tasking cause your projects to take longer? The second step is deciding whether it’s worth fixing. Do you know your projects’ cost of delay (even approximately – is it $1K, $10K, $100K per week)? Can you calculate a rough cost of multitasking in your part of the organisation?