The Unintended Consequences of Individual Metrics

Several years ago when I was working with an e-commerce company we came up with a framework for how to grow transactional shopping revenue. We called it "the Box".

The idea was to get shoppers into the box (acquire new users and get them to come back to our sites regularly). And then, once they were in "the box", to make good things happen (get them to buy lots of stuff).

We setup two separate teams: one team was focused on driving traffic and the other was focused on converting that traffic into dollars.

The "traffic driving" team didn't worry about shopping conversions and the "conversion" team didn't worry about driving traffic. We put the teams in silos and told them to focus on their goals. The thinking was that if both teams did their job, overall revenue would grow.

The beauty of the framework was that when weekly revenue grew, we could very easily determine who deserved credit. Was the increase caused by something that the traffic driving team did or something that the conversion team did?  Very rarely was it both. Neither team could hide behind another's success. We could easily identify the initiatives that we're contributing to overall revenue and those that weren't. It seemed like a great model.

But we quickly saw that the structure we setup caused some problems.

The traffic driving team, in an effort to drive traffic (as opposed to revenue), found some quick, easy and suboptimal ways to drive traffic to our sites. For example, they'd email millions of users with an offer from a high-end car company. The response rate would be great and traffic grew, but nobody bought (low conversions). Users just clicked around and looked at the cars because they were interesting. Most weren't planning to buy, or if they were they were planning to buy offline.

At the same time, the conversion team put brands that converted well (such as Target and Wal-Mart) front and center on our websites. While those brands did convert well, they produced small average order sizes and didn't pay us a significant commission. Combine that with the fact that the traffic driving team wasn't pushing shoppers to the offers that the conversion team was promoting and we quickly found that our user experience was disjointed.

It became clear that we couldn't have our teams operating in silos. To maximize revenue, they had to work together. They had to collaborate. They had to do more than just their own job.

This initiative underscores the challenges around siloed teams and metrics. When high performing people are given clear objectives with quantifiable metrics attached to them, they'll very often accomplish those objectives. And there will very often be unintended consequences from that accomplishment.

All of that said, even after that experience, I still strongly believe that managers must create clear, measurable metrics for every employee in the organization that only that employee can control. It's a crucial part of an accountable, high performing organization.

But at the same time managers need to closely monitor whether or not those siloed metrics are positively or negatively impacting the overall health of the business. Setting up a framework where individuals and teams are focused on producing impactful work week to week is the (relatively) easy part. Getting multiple teams focused on snyergistic activities that add to the overall value of the business is much more difficult. And much more important.