Trust Employees

You Can Trust Employees with Peer Bonuses. Here’s the Data.

By Raphael Crawford-Marks on January 06, 2015

Can you trust employees to give peer bonuses responsibly?

The idea of employees having control over recognition and bonuses is sometimes regarded with skepticism. In fact, some cite Bonusly's high engagement numbers (90% of employees giving bonuses each month) as a sign that there might be something nefarious going on: quid pro quo bonuses.

After analyzing the behavior of our entire user base, it turns out that this fear is not supported by the data. In fact, you can definitely trust employees with peer-to-peer bonuses, as they tend to avoid giving bonuses to those who recently gave them one.

What we talk about when we talk about love reciprocity

When two users give each other bonuses within a short period of time, this may look like reciprocity (“the practice of exchanging things with others for mutual benefit”). We generally want to discourage that, because bonuses ought to be given to celebrate great work that drives value for the company.

Skeptics often assume that if a lot of employees are giving and receiving bonuses, then the only possible explanation is reciprocal giving (A gives to B and B gives to A). This assumption is incorrect - it is possible to have high participation without excessive reciprocity:

High participation with no reciprocity. Trust employees. 100% participation with no reciprocity.


In the above example, 100% of employees both gave and received a bonus. Yet there was no reciprocation.

It’s also important to note that the appearance of reciprocity can happen for perfectly legitimate reasons.

A gives to B and B gives to A. This can be very good. Trust employees.A gives to B and B gives to A. This can be very good.


Let’s say my colleague Jane wrote a popular whitepaper that resulted in a huge jump in inbound leads. I give her a bonus for that. A few days later, I provide her with some key data and help her work it into a blog post. She gives me a bonus. In this case, A has given to B and B has given to A, but the recognition given was specific, honest, timely and appropriate.

As a manager or CEO, you want this kind of thing to happen. The behavior being recognized is generating significant value for the company. And it’s motivating to employees to know that their work is recognized. So our goal isn’t zero reciprocity - we just don’t want excessive reciprocity, which could be an indication of abuse or cheating.

The question is, how do we measure reciprocity? And what rate of reciprocity is good?

We dove into our data to answer these questions.

Neighborhoods and Expected Reciprocity

If you work at a company larger than ~10 people, do you interact with everyone? Almost certainly not. Within a company, you have a “neighborhood” of colleagues - the people on your team, in your department, etc.

We looked at our data, and found that on average, employees interact with 6 or 7 people in a 3-month period (6.6 to be exact). This varies a bit from company to company, but interestingly, there is no correlation with company size. An employee’s neighborhood is not likely to be any bigger if they’re at a 500-person company than if they’re at a 50-person company. 

Average Neighborhood Size
On average, employees exchange bonuses with a neighborhood of ~7 employees, regardless of company size.
Knowing what neighborhoods look like, we can estimate what a normal reciprocation rate should be. On average, Bonusly users give 1.3 bonuses per week (about 5 per month). This means that anytime I give a bonus to someone in my neighborhood, there’s a good chance the recipient gave one bonus within the last week. And there’s a 1-in-7 chance that they gave that bonus to me.

Our expected reciprocity rate is the expected frequency that reciprocal bonuses would occur if employees gave bonuses randomly to colleagues in their neighborhood. Given the neighborhood sizes and give rates for companies using Bonusly, we would expect about 17.1% of bonuses to appear reciprocal.

Employees Avoid Peer Bonus Reciprocation

After analyzing the behavior of our entire user base, it turns out that employees are actually a bit averse to reciprocation: they tend to avoid giving bonuses to those who recently gave them one.

 Peer Bonus Reciprocity

Incidence of reciprocity in Bonusly is lower than one would expect.

On average, the incidence of reciprocation was about six percentage points lower than what we would expect from random chance.

Incidence of Reciprocity \
The vast majority of Bonusly companies have lower-than-expected reciprocity.


The vast majority of companies using Bonusly have lower-than-expected reciprocity, some of them significantly so. For those with higher-than-expected reciprocity, the discrepancy is slight (less than 10%).

 Incidence of Reciprocity 2

Lower-than-expected reciprocity holds true across company sizes, as well.

Conclusion: Trust Employees

While this result is great evidence that employees are using Bonusly responsibly, I think the result has more profound implications for how we should treat our employees.

The assumption that employees are intentionally cheating (at worst) or unconsciously giving quid pro quo bonuses (at best) is indicative of a very common fear: that employees can’t be trusted to use their own judgment.

This fear manifests itself in many ways:

  • “Peer bonuses will never work.”
  • “Transparency will never work.”
  • “Distributed teams will never work.”
  • “Flexible hours will never work.”
  • “Flat teams will never work.”

Our data on peer bonuses raises the possibility that many of these assumptions may be wrong. At the very least, they ought to be explored with an open mind, instead of dismissed out of hand.

Written by Raphael Crawford-Marks

Raphael Crawford-Marks

Raphael is co-founder and CEO of Bonusly, where his focus is on creating a work environment that enables the Bonusly team to thrive.