- Labor productivity metrics are an outdated, counterproductive way of measuring human impact in a world where humans are focused on producing ideas (vs. widgets).
- Using output over time to measure a knowledge worker’s performance ignores the quality of that output and the results it led to.
- Instead of productivity, we should be measuring whether knowledge workers are achieving the right results for customers and whether they’re happy and healthy while doing it.
You may have heard the phrase “what gets measured gets managed.” So when we want something done, we slap a metric on it and manage toward that metric. Seems reasonable enough, right? It’s not, though. It could actually be one of the worst strategies ever. Let me illustrate with a story a colleague shared with me.
A director at a software company was freaking out because development was weeks behind schedule. So he did what any thankless brute would do: he instituted mandatory Saturdays.
For the developers, that is. Not for him.
He would drive past the office on Saturdays to check the parking lot and, seeing his developers’ cars there, would roll away satisfied. Naturally, his team grew a wee bit resentful. So instead of actually coming into work on Saturdays, they started having friends pick them up from the office on Friday afternoons and taking cabs in on Mondays.
Lesson: if cars in the parking lot are what the big boss is measuring, then cars in the parking lot he shall get!
I probably don’t need to tell you that when the product ultimately shipped, it was months late.
The productivity paradox
We encounter the same measurement/management problem when using productivity as a metric for team and individual performance. See, productivity is just a mathematical equation: output divided by time. This has two implications:
- When we talk about productivity, we are inherently and inescapably talking about output – not outcomes.
- When we talk about increasing productivity, we’re really talking about increasing output.
Trouble is, more output doesn’t necessarily mean better results. As best-selling author Dan Pink told me recently, he could write two mediocre books in the same time it takes to write one really good book. Two books is twice the output! Twice the productivity! Hallelujah! But his publisher would have some pretty choice words for him because mediocre books don’t sell. So even though he was twice as productive, the results would be crap.
And yet, as a society of knowledge workers, we are obsessed with productivity. We’ll click on any article with that word in the headline. We “ooh” and “ahh” over reports that productivity increased during the pandemic when everyone was working from home (which I dispute, by the way – more on that later).
So, how did this happen? Is it doing us any harm? (Spoiler alert: yes.) And if we’d do ourselves a great service by killing productivity as a performance measure, what the devil do we replace it with?
How did we get here?
The way we think about productivity is based on a 250-year-old construct. It was good in its time, especially in an agricultural or manufacturing context. Technological advances like threshing machines and mechanized looms increased the daily output of our farms and factories by orders of magnitude. And because we didn’t scale back the number of hours worked, productivity increased.
As the industrial revolution progressed, productivity as an obsession metric became more and more entrenched. Further advances in farming equipment meant less human labor input was required to feed us all, freeing up more people to work in the factories that were taking the human drudgery out of making stuff like clothes and furniture. Mining and logging kicked into high gear producing the raw materials to fuel it all. Output per day was still a salient measure of performance for the vast majority of the economy.
By the time agriculture and industry became so mechanized that humans transitioned to knowledge work – office-based jobs, education, health care – the obsession with productivity was woven so tightly into the fabric of society that measuring efficiency any other way seemed like cheating. We worshipped science! Because progress! Mathematically calculated productivity metrics give us hard data, and data feels like certainty. Anything else is just smoke and mirrors.
The cult of productivity is counterproductive
Fast-forward to now, and the production of widgets is almost exclusively the domain of machines and robots. But the production of ideas is exclusively human. Knowledge workers are incentivized to use their heads and hearts as much as their hands. Tools like email, chat, word processors, and printers make it faster to communicate and act on our ideas – and productivity can measure that impact. But tools have very little effect on the quality of our ideas – which means productivity can’t measure our creativity.
Productivity also can’t measure whether the work we put out there is the right work. (Remember Mr. Pink and the problem with writing two mediocre books?) A developer might fix 10 bugs today, but if these are trivial bugs or ones that customers rarely encounter, their work isn’t having a material impact on the product or the business.
Yet, chances are, this developer’s team is judged by how long it takes to ship stuff, be it features or fixes. This leads to perverse incentives. If the team is rewarded for shipping quickly, they’re incentivized to ship a bunch of lightweight features that may not noticeably improve the user experience and therefore do nothing to make the product more attractive. They may also be rewarded for catching defects early because dealing with them later slows down new development. But that has a similarly unintended consequence. It means they’re incentivized against taking on ambitious projects that require deeper, more structural (and riskier) work.
And then there’s the effect on our work-life balance. Unless you’re giving people new tools, pushing them to be more productive means pushing them to work longer hours and/or rush through their work, both of which increase stress levels. As leaders, we’re shooting ourselves in the foot measuring employee productivity this way, because our team members can’t think creatively or bring their full abilities to bear on a problem if their brains are too busy feeling anxious and exhausted.
Better alternatives to measuring productivity
Productivity has always been a good way to measure the impact of machines and capital. It’s just never been a good way to measure the impact of humans. So what metric should we use instead? How do we reframe how our input is valued? How do teams and companies evolve their culture to measure more meaningful things? How do we shift from focusing on efficiency to focusing on effectiveness?
Emphasize outcomes over outputs
At a high level, we need to emphasize outcomes for our customers and/or business and de-emphasize our output of effort. Instead of telling IT admins to set up 10 new load balancers this quarter, we should tell them to improve site performance by 10 points. Instead of telling a marketer to publish five blog posts, tell them to increase web traffic by five percent.
One beauty of shifting to an outcomes mindset is that it not only keeps us focused on results, but also frees us up to innovate in the pursuit of those results. There are loads of ways to improve system performance or drive more traffic to a website. But once we say “write five blogs,” we’ve significantly reduced the opportunity for creativity. So it’s critical to articulate goals as the results we’re after, not as to-do lists, then let the people doing the work decide the best way to go about it.
Another advantage is that an outcomes mindset encourages an ownership mindset. When we’re focused on productivity, we manage toward deadlines. Once the work has shipped, we clap the dust off our hands, congratulate ourselves on getting it sorted quickly, hope it achieves the result we were after, and then never think about it again. By contrast, managing toward outcomes leads us to ship some work, gather feedback on it, and iterate as many times as we need to until the result is actually, verifiably achieved. As execs at both Hubspot and Twilio have noted, this encourages every employee to think like an entrepreneur.
Our teams are defined by three things: the customer they’re serving, the mission they’re on in service of that customer, and the metrics that tell us whether we’re doing a good job.– Jeff Lawson, CEO at Twilio
Pay attention to leading indicators
There’s an exercise called Goals, Signals, and Measures that I’ve run with dozens of teams. For any goal (outcome), you determine what signals will indicate that you’re on the right path and what measures will confirm that the goal has been met. We need a similar method for measuring a team’s effectiveness, because outcomes are lagging indicators; we need some early signals to listen for. Here are a few ideas:
- Map out (and celebrate) milestones. There’s nothing wrong with paying attention to how long it takes to achieve an outcome. Being able to say “We’re halfway to our goal of reducing customer support calls by 30 percent this year!” is great for morale – especially if you got there faster than expected. Likewise, knowing that you’re running behind generates urgency, which often leads to innovative thinking as you figure out how to get back on track. And, by the way, tracking milestones also forces you to build measurement into your work from the get-go, which helps you communicate productivity growth to stakeholders in a way that feels concrete.
- Build feedback loops. If an IT team wants to free up more time for special projects, they might set up a knowledge base so people can solve everyday issues themselves instead of raising a help desk ticket. The team can then look at how often the knowledge base articles are accessed and how often tickets are raised as signals that their strategy is working (or not). But feedback loops aren’t just for technical teams anymore. A publisher that wants to build a bigger audience might look at how often their content is shared on social media. A finance team that wants to improve its models might look at the delta between forecasted and actual results. And any team can go out and actually talk to customers. It’s scary, I know. But it works.
- Look for continuous improvement. Measuring improvement in this context generally means looking at rates of growth or reduction. Are service outages decreasing over time? Is the rate of new customer acquisition staying steady or improving year-over-year, even as maintaining that growth gets harder as the baseline becomes larger? If your rate of change is improving, that indicates you’re learning as you go. On the qualitative side, you can hold retrospectives to reflect on recent work, pluck out some lessons learned, and apply them to your work going forward.
- Value employee well-being. Health and happiness aren’t byproducts of efficiency. Rather, they are essential ingredients for being effective, as I noted earlier. As leading indicators, you might look at the number of planned vs. unplanned days off. If a team member has far more unplanned PTO days, that might be a sign of burnout or other personal issues they’re navigating. Consistently working more than 40 hours per week could be another sign. Either situation should prompt a conversation during your next 1-on-1. Make it about the person instead of their projects and ask what sort of support they need.
If leaders took on the task of retiring output-based measurements of success and instead focused on outcomes, what would that list of success metrics look like? It’s not that we need to get rid of quantitative measures. We just need more meaningful ones.
The best time to kill the cult of productivity was 20 years ago. The second best time is today. Let’s get after it.
Get stories like this in your inbox