I want a number, a metric, that tells me how productive our teams are!”
challenged my former Head of IT, some years ago. Certainly, it’s a reasonable request to ask how productive a team (or a whole system) is.
But first let’s look at the why behind the question. Why do we measure productivity? Because we should? Because we can? For transparency? Accountability? To drive behaviors and influence culture? Should we measure it at all?
There are three simple, impactful metrics (for agile or waterfall workplaces) that I informally collect (through conversations is a good start) to quickly gauge productivity and how healthy and high-performing an organization is.
Lead time is the queen of metrics. Lead time tells me how long it takes work to travel through the whole system, from concept to cash.
How quickly is value delivered to customers? Long lead times indicate waste. Lean experts Mary and Tom Poppendieck describe how over 80 percent of time-to-market can be “waste” in the form of delays and non-value-added work. That quickly snowballs into a double-pronged, multi-million dollar cost of delay that directly hits a company’s profits via the delay in value delivered to customers and thousands of wasted person hours.
What do long lead times look like in the real world?
This week I’m working with the largest company of its type in this state. It regularly takes up to six months to get a business case approved, and another 12 months to deliver a project. That’s an 18-month lead time to deliver value to customers. Given that requirements change at an average rate of around 2 percent per month (based on Capers Jones’ empirical research in the 20th century, and it’s probably higher in 2016,) this means a project that goes live today is delivering requirements that were signed off on 12 months ago and have changed (degraded?) by over 24 percent.
This company’s volume of work is expected to increase at least 30 percent in the near future (with no headcount increase.) What happens when we add 30 percent greater volume to an already chockablock freeway? It reduces our speed by an order of magnitude.
This company is adding risk to its portfolio by having such long lead times. Are the teams productive? Not nearly as productive they could be. What actions should they take to reduce lead times? Just reducing the batch size of work (e.g. from 12-month projects into small, discrete features) and setting work in progress (WiP) limits will often double throughput (i.e. halve lead time) as described by Lean management guru Don Reinertsen. These are things you can start doing sooner rather than later.
But, by itself, lead time doesn’t tell me how productive a team is.
Predictability complements lead time and has an equal seat at the head of the table as the king of metrics. Not only do I want a short lead time, I want to reliably know when work will be done and when value will be delivered to customers. Predictability is not boring—it’s the new black. And it’s sure better than 50 shades of grey, so to speak, in terms of guessing when something might be delivered.
The city I’m working in suffered floods not so long ago. I asked my client, whose offices overlook the river, whether the council knows the volume of water in the river and its rate of flow, i.e. how much water flows into the nearby sea every day. “Of course,” he replied. “So, what about your portfolio? What volume of work can it handle and how quickly will that work flow out to customers?” My client didn’t even pause.
We don’t know. We don’t really know what our capacity is at the portfolio level or how quickly we can deliver work.”
That’s not unusual in this type of organization. It would be unusual in manufacturing, where every widget is a physical item and easily traceable. But where work is less tangible it’s easy for “invisible waste” to significantly erode capacity.
Predictable delivery not only increases profits and reduces bottlenecks, it has a more important outcome: it creates trust, trust that teams will deliver on time and that the portfolio can and will deliver the number of features (or requirements) promised. I give my business to companies I can trust—that deliver when they say they will—over companies that don’t deliver when they say they will.
What actions can you take to increase predictability? You need to know the capacity and velocity of your portfolio. Once your requirements are logically grouped into features (see above,) use relative sizing (starting with a small-ish and well-understood feature) to quickly get a view of how much work is in-flight and in the pipeline.
T-shirt sizing is fine if your stakeholders are new to story points (which you can later map over the t-shirt sizes.) It will probably be “way too much” work, which is where prioritization comes in (a topic for another time.) Then, populate “just enough” features to be assigned to the next program increment (say, 12 weeks.) And do this activity with the people close to the work, not far-removed stakeholders.
When I find a company (or a team) with short lead times and high predictability, it’s a good indication that it is productive (although it doesn’t tell me that they are delivering the right things—another topic for another time.) But there’s one other metric that trumps both lead times and predictability.
Happiness is the most important metric because in a knowledge economy, talented people are the competitive advantage. Are our people (and customers) happy? Simplistically, happy employees deliver good products, which lead to happy customers and good profits. And, the reverse is usually true: an unhappy employee is more likely to deliver a poorer product, leading to unhappy customers and poorer profits. "People, products and profits—in that order,” as our own CA Agile Business Unit GM, Angela Tucci, reiterates. I want to know if my employees are happy or unhappy and why, because it’s closely linked to motivation. As Dan Pink’s now cult classic video explains, give your people autonomy, mastery and purpose and they will be motivated to change the world.
How do we find out whether our people are happy? Ask. Not (just) via an anonymous, annual, online tick box survey. Ask via team retrospectives. Ask via one-on-one or small group sessions. Use a simple 1-5 Likert scale if you want an easy way to quantify the qualitative data. Ask what’s making people happy and unhappy. Frequently improving what’s making people and teams unhappy improves our other two metrics: lead time and predictability.
For example, my client is generally happy but is anxious because the organisation needs to pull 30 percent more work through the “system” as part of its growth objectives. My client’s teams perform reasonably well but are frustrated because there are bottlenecks around key roles and these delays generate significant non-value-added workarounds. Improving these problems would make these people happier and improve lead times and predictability, and lead to happier customers and greater profits.
Let’s return to my former Head of IT and the quest for a single metric for productivity: this may be a holy grail for another explorer. But, armed with metrics for lead time, predictability and happiness, I can reasonably and efficiently infer sustainable productivity—not only at a team level, but at a portfolio and company level.
And so can you.
This blog is syndicated from CA Technologies. Read more on Highlight, the CA blog.