Why You Should Be Careful With People Recommendations

Recommendations are often no better than a Google search.

They say you find the best recruits, service providers, and investors by asking for recommendations. So I decided to test the theory.

I listed all my service providers and partners over the last three years, along with how I found them. Surprisingly, third-parties that came via a recommendation were less likely to succeed than those that came from other sources.

In fact, some of the worst people I’ve worked with came via my network.

What’s going on here?

Not all recommendations are created equal

Why do some recommendations work out, while others fail? Looking at the data, I discovered that one of the following was true for all recommendations that turned out badly:

1. The referrer hadn’t worked directly with the recommendation for a long time — or not at all. For example, they knew them from the gym, or were the partner of their best friend.
2. The referrer didn’t have credibility in the area of the recommendation — for example, someone with little-to-no marketing experience recommending a marketer with whom they worked for a day.
3. The referrer was someone I didn’t know well or fully trust. For example, I knew them from an online community, or I wouldn’t have wanted to work with them personally.

On the other hand, the fantastic recommendations came from people who matched this description:

Someone I think is awesome that recommends someone they think is awesome after working closely together.

That’s why, when I ask my network for a recommendation, I seek answers to these questions:

• What’s your experience in this area?
• How long have you worked closely with this person/company?
• Did you work together directly?
• Are you working together now? If not, why not?

I then ask myself if the referrer is exceptional at what they do. If they aren’t, even if the questions check out, I proceed with extreme caution.

Don’t trust people, trust algorithms

But I still needed to understand why third-parties who weren’t recommended did so well.

Daniel Kahneman is a Nobel laureate who knows a lot about making decisions, and he offers some insightful advice when it comes to making them:

When there’s the possibility of using an algorithm, people should use it. We have the idea that it is very complicated to design an algorithm. An algorithm is a rule. You can just construct rules.’

This is why structured interviews beat gut-feel ones — the process of asking the same questions to each candidate and rating them on pre-selected criteria helps override the biases of intuition.

It also helps explain why the people I found via a careful process outperformed the recommendations. The social proof associated with the recommendations was too powerful to override and I trusted gut-feeling too much.

Ultimately, recommendations aren’t a silver bullet — and they may even be a blindspot. They need to be carefully assessed, and if you want great results, you should run a structured process too.