The other day I tweeted, “Every critical metric at a company needs a counter-metric to prevent the wrong kind of optimization.” It got a reaction and I thought to expand on the topic.
Have you seen any of these examples of metrics + reward systems gone awry?
- The marketing and demandgen teams are tasked with bringing on leads, so they ramp up paid acquisition and low-quality/high-velocity content. They get tons of leads, but the quality is poor and the sales team wastes a huge amount of time on bad prospects.
- The engineering team is told to ramp its throughput of story points. They decide to hit their new target without gaming the system, but it leads to a huge jump in half-baked features and architectural decisions, not to mention team burnout.
- The customer success team is tasked with increasing a key product metric, so they start doing the task on behalf of their customers. They hit their metric, but an entire generation of customers is trained to expect free labor and the company’s unit economics are wrecked.
- The content strategy team is measured on traffic figures, so they steer into sensationalist click-bait, driving up the numbers with people who don’t care about the product and don’t stick around for more than six seconds.
- A double-sided marketplace sets aggressive metrics on growing the demand side of their business. They enact a new set of actions and policies that over-optimize for their demand-side customers, anger the supply side and thus spawn a negative spiral.
- The head of sales is tasked with hitting big numbers, so his reps over-sell on the product’s capabilities, stretch the ICP beyond recognition, and commit to custom feature requests. Twelve months later, after paying out big commissions, the churn rate skyrockets leading to a board panic and ensuing layoffs.
- The CEO’s comp is tied to the stock price, leading to short-term thinking and business decisions designed to please Wall Street, undermining the foundation of the business.
Metrics are an essential management tool, but boy-oh-boy have I seen them cause havoc because people failed to think through the side effects and ramifications. That’s why I strongly believe that every numerical goal needs at least one counter-metric to watch.
Questions to Ask
When you pick an ambitious target to hit, you must think through two related questions:
- What are the negative shortcuts, hacks and/or gaming that we need to protect against?
- What are the possible negative consequences and things we might neglect in relentless pursuit of the target?
You should be able to come up with one to three (ideally no more) important counter or health metrics to keep an eye on, as you chase your goal.
I am an unabashed fan of Christina Wodtke’s approach to OKRs in her book Radical Focus. For one thing, she recommends that the OKR review process stay forward-looking, not backwards-looking. But equally important, in her OKR quadrant framework, she has an important box that few people seem to understand at first glance: health metrics. These are the things you want to protect while you chase your ambitious key result.
I’ve seen people go wrong with this box by using it as a dumping ground for any metric they also want to track. But really, to me, the health quadrant should be those essential counter-metrics you uncovered when asking and answering the above two questions.
To sum up, picking and aligning those ambitious OKRs is only part of the process you need to go through, especially if you are going to tie them to compensation or reward mechanisms. It’s also essential to think through and design for the side effects.