Why Celebrating Small Wins Isn't Enough: The Science of Ritual

I had the suspension notice open in one tab. In the other, I was deleting Friday.

Friday is a standing dinner I’ve kept for months. It’s a small table where we mark the week, whatever the week turned out to be. Cancelling it felt obvious. The account I’d spent a quarter building had just been locked for the second time in nine weeks, and marking the week seemed like a strange way to answer a week that had failed.

My cursor was already on the delete button when I caught what I was doing. I was treating celebration as a reward. That’s the default assumption, and I think it’s wrong. Most of what gets written about the importance of celebration ritual ends up in the same place: celebrate your wins, especially the small ones. Follow that advice and celebration turns into a payout. Payouts stop when the wins stop, and that’s exactly when you need one.

So here’s my claim: celebration is infrastructure. Think of a road. You build it before the traffic shows up, and you maintain it whether or not anyone drives on it this week. If you let it go, you find out at the worst possible moment.

Isn’t “Celebrate Small Wins” Enough?

No. That advice ties celebration to output, so it disappears during a losing streak, which is when it would do the most good.

The advice usually traces back to Teresa Amabile and Steven Kramer’s research on the “progress principle.” They analysed nearly 12,000 daily diary entries from 238 knowledge workers across seven companies and found that people’s best days at work were the ones where they made progress on meaningful work. Their 2011 Harvard Business Review piece, “The Power of Small Wins”, also reported a survey that asked 669 managers to rank what motivates employees. Progress came last. Only 35 of the 669 put it first.

The finding holds up. The way people usually read it is backwards. Amabile’s point was that managers don’t see progress, so it goes unmarked. What people took away was “notice the wins and throw a party.” But noticing is perception, and celebration is a separate mechanism. Hook celebration to wins and it picks up all the variance in your output. The founder with a flat month gets nothing. So does the writer staring at a dead draft, and so does the team a week after a failed launch. Those are the people whose motivation is lowest.

Infrastructure doesn’t check your metrics first. It runs because it’s Friday.

Doesn’t Celebrating Make You Complacent?

The evidence points the other way. Shared rituals of acknowledgment tend to come before better performance.

This is the strongest objection, and the stoic-productivity crowd means it. The argument goes that celebrating releases tension, released tension kills drive, and drive is what gets results. It sounds rigorous. It doesn’t hold up against the data.

In 2010, Michael Kraus, Cassey Huang and Dacher Keltner coded every fist bump, high five, chest bump and team huddle in NBA games from the start of the 2008–09 season. Their paper in Emotion, “Tactile communication, cooperation, and performance”, found that teams that touched more early in the season performed better later on, both individually and as teams. That held after controlling for player salary, preseason expectations and early-season performance. The high five was doing coordination work, and it predicted cooperation and wins months down the line.

What breeds complacency is ending effort. Marking it is something else. A celebration that closes a project (“we’re done, relax”) can make people coast. One that runs on a fixed schedule (“it’s Friday, let’s look at the week”) doesn’t close anything. It’s a checkpoint, and the work carries on afterward. The complacency argument only works if celebration is a finish line, and roads don’t have finish lines.

Isn’t Ritual Just Superstition?

No. Ritual changes measurable physiology and behaviour even when the actions themselves are arbitrary. You don’t have to believe in anything for it to work.

Take the most cynical version of the objection: the gestures are meaningless, so any effect has to be placebo. Fine. Look at what the experiments actually did.

Alison Wood Brooks and colleagues had participants sing Journey’s “Don’t Stop Believin’” in front of a stranger, which is about as brutal an anxiety test as a lab can ethically run. Some participants first went through a short, pointless routine: draw what you’re feeling, sprinkle salt on it, count to five, crumple the paper, throw it away. Their 2016 paper in Organizational Behavior and Human Decision Processes, “Don’t stop believing: Rituals improve performance by decreasing anxiety”, found that the ritual lowered self-reported anxiety and heart rate and made the singing more accurate. The detail I care about is that the effect depended on calling it a ritual. When the same actions were described as “random behaviours,” they did nothing.

Kathleen Vohs, Yajin Wang, Francesca Gino and Michael Norton found the same pattern with chocolate. In their 2013 Psychological Science study, “Rituals enhance consumption”, people did a short ritual before eating a chocolate bar: break it in half, unwrap one half, eat it, then eat the other. They rated the chocolate as more flavourful, savoured it more and said they’d pay more for it. But they had to do the ritual themselves. Watching someone else do it had no effect.

So the frame is the mechanism. Structure you’ve named and repeated changes how the next thing feels. Calling that superstition makes about as much sense as calling scaffolding decoration.

The Vohs result has one more implication that the celebrate-your-wins advice misses. The ritual made the chocolate better, and it didn’t need good chocolate to start with. If you attach a ritual only to good weeks, you’ve spent it on the weeks that needed it least.

What If There Is Nothing to Celebrate?

Then the ritual matters more. Structured ritual after a loss measurably reduces grief and gives people back a sense of control.

Here the infrastructure model and the reward model split completely. A reward system has nothing to offer a bad week. Infrastructure carries whatever load shows up.

Michael Norton and Francesca Gino studied this directly. In their 2014 paper in the Journal of Experimental Psychology: General, “Rituals alleviate grieving for loved ones, lovers, and lotteries”, people who performed rituals after a loss reported less grief than people who didn’t. The losses included a death, a breakup and even losing a lab lottery. The mechanism was restored control. A loss takes agency away, and a ritual gives a little of it back through a sequence you chose to carry out.

Anthropology shows the same thing at the scale of a whole community. San Pedro Manrique, a Spanish village, holds a fire-walking ritual every year. Ivana Konvalinka and colleagues put heart-rate monitors on the fire-walkers and on the spectators. Their 2011 PNAS paper found that during the ritual, the fire-walkers’ heart rates fell into sync with the heart rates of spectators who were related to them. Unrelated spectators didn’t sync up. People who share a ritual end up sharing a pulse.

That’s why I didn’t delete Friday. The suspension was what the week contained. The dinner holds whatever the week contained. The mistake is assuming it’s only meant for good weeks.

A celebration that only runs after wins is a reward system. Reward systems break under exactly the conditions they’re supposed to survive.

Why Do Forced Team Celebrations Feel So Fake?

They’re imposed instead of performed, and they mark events someone else picked. Both of those break the mechanism the research points to.

This objection has the most real-world evidence behind it. Everyone has sat through the mandatory pizza party after a round of layoffs, and everyone has watched a wins channel fill up with emoji nobody means. If celebration is infrastructure, why does so much of it feel like theatre?

The studies already answer that. Vohs found that watching a ritual did nothing and doing it did something. At a pizza party, management runs the event and staff sit and watch. Brooks found that the frame mattered. Once everyone privately calls a celebration hollow, it has already been relabelled as “random behaviours,” and that was the version that failed.

The third reason comes from Shelly Gable’s work on capitalization, which is her term for what happens when you share good news with someone. Her 2004 paper in the Journal of Personality and Social Psychology, “What do you do when things go right?”, found that telling other people about good events was linked to more positive affect and well-being, beyond the effect of the event itself. How the listener responded made the difference. Active, engaged responses strengthened relationships. Passive or dismissive ones didn’t help.

A mandatory celebration produces passive responses by design. Nobody chose to be there, so nobody really engages. It fails because it’s missing the parts that make ritual work: people doing it themselves, a frame they believe, and real responses to each other. Those are construction requirements. Mood has nothing to do with it.

What Does Celebration Infrastructure Actually Look Like?

It happens at a fixed time, the people in it do something, and it doesn’t depend on outcomes. It runs on schedule and holds whatever the week produced.

Here’s what the studies above add up to:

• It’s scheduled. It runs weekly or at the end of a cycle and never waits for a milestone. The NBA touches happened every game, and Amabile’s diaries were daily. Frequency matters more than size.
• Everyone takes part. Each person names one thing that moved, one thing that broke and one thing they’re carrying into next week. The Vohs effect depended on doing it yourself.
• It’s called a ritual. Brooks’s participants only benefited when the actions were framed that way.
• It works in a bad week. If the format can only hold good news, what you’ve built is a trophy shelf. Norton and Gino’s results came from rituals after losses.
• People respond. When someone shares something, the others actually engage with it. Gable’s active, engaged response is the load-bearing part.

None of it costs much. My Friday dinner is a table, two hours and one question: what happened this week? On 25 September the answer was that the account had been suspended again, one week after I raised its reply volume. We talked about it anyway, and that was the point of having the dinner.

What Changes If Celebration Is Infrastructure?

If I’m right, people stop asking “have we earned a celebration?” and start asking “is the ritual still running?”

Managers would stop budgeting celebration as a morale expense tied to quarterly results and start treating it as maintenance. You don’t cut the road budget in a recession, because that’s when the road carries the most weight. (Plenty of people cut it anyway, which is roughly my point.) Founders would stop cancelling the Friday table during a flat month, because flat months are what the table is for. Teams would stop judging their culture by whether people enjoyed the party and start checking whether the ritual survived the last failure.

The standard advice treats celebration as dessert, served once the work is done. The research describes something closer to a load-bearing wall. It coordinates teams months in advance, steadies your nerves before a performance and gives back some control after a loss. You can’t add that wall once the building has started to sway.

Celebration you build only for good days is decoration. Celebration you build for every day is still standing when the week goes wrong, and that’s usually the only time anyone checks.