SAFe Value Stream Mapping

As in Orwell’s dystopian 1984, SAFe reinvents terms to suit its needs and hijack words but not their underlying concepts. One of these terms is value stream. The reason is evident. If you look at SAFe through the lens of traditional Value Stream Mapping, you’d quickly realise that most of the time, no value is being generated. Like inventory sitting in a warehouse, the output may have value, but at what cost and when will it see the light of day?

There is so much overhead that might arguably protect value, but it does not generate any value. In a traditional value stream exercise, one looks at all of the activities and wait states for a process from end to end. Now consider a standard SAFe process—even without all of the portfolio over-head, none of which generates value.

Keep in mind that part of the promise of SAFe is not only that it’s Agile (or agile)—it’s neither—, but that it’s scalable. Follow the money, and you’ll see that NONE of the activity outside of the essential picture even has the potential to generate value. It’s all deadwood activity.

Don’t get me wrong. I am not claiming that there is no value protection to be found. I’m talking about value generating activity. Let’s begin with these essentials and build from there with a focus on value generation and wait states (perhaps, we’ll call them waste states).

We start with an idea. Let’s pretend it actually came from customers. More likely, some product owner or committee dreamt it up. No value generation here. By definition, any analysis is not generating value, so let’s keep moving. Let’s create a backlog and groom it. No value to be found here. Let’s plan the work. Still no value. Let’s create sprints or use Kanban. It doesn’t matter which. Neither is generating value. Let’s start a sprint. Finally, some potentially value-producing activity. Sure, it’s interspersed with more analysis and coordination, but we know we can find measurable progress. Of course, there is dev testing and unit testing. No value being generated here. We may have actually seen the end of value generation, but we’ve got a ways to go before we actually convert that value to cash. Is there a QA handoff? No value there. What about UAT? Integration testing? Nope. Just preventing leakage. Are we ready to release yet? No? Just wait? OK. Go!

What percentage of this process generated value? Five per cent? Ten? Now let’s scale up.

In our large solution, none of the activity generates value. It’s all over-head. So, what about the portfolio view?

Nope. Just more deadwood. All cost. No value. But wait, what about this ‘business agility’? These are empty words. The business was already present and there is still no agility. Obviously, if I go all in and add Large and Portfolio, I am adding zero and zero to whatever paltry value base I had at the start—only I’ve diluted it with one or two layers of overhead. In maths, we’ve increased the denominator and held the numerator steady.

Am I saying that this portfolio layer is worthless? No. Not at all. But I am saying that it generates zero value and at a cost. Is it worth it? It might be, but that doesn’t mean it generates value.

But SAFe says it has value streams? What gives? As already mentioned, SAFe redefines what a value stream is. Probably to redirect the unwary. In their parlance, a value stream is essentially a workstream without regard to whether it delivers value or even has the potential of delivering value. If you aren’t measuring value, you won’t know the difference anyway.

In summary, from a value stream mapping perspective, SAFe creates a slew of wait states. Unless you are fermenting wine and ageing cheese, wait states add no value—just cost.

Now that we’ve walked through this, you should realise that you can manage a delivery portfolio without all of the pretense of SAFe, and you should. SAFe is a dog with fleas. Ditch SAFe, and just implement via old-school PPM portfolio management. And before you imagine that I am saying that you should adopt another scalable Agile framework, perish the thought. Agile is inherently not scalable. Scalable Agile is a sucker’s bet. Don’t be a pigeon. Full stop.