Steam 愿望单膨胀真实存在，但依然需要它们

This is a deep dive into the results of the Steam wishlist and sales data I have been collecting from readers. The survey is still active so if you would like to contribute, please complete this form:

Official HTMAG Steam wishlist and sales survey 

If you want to get a lot of views on YouTube, an easy trick is you take a nugget of common knowledge and say “<nugget> doesn’t matter anymore, <Y> is the true marker of success.” Boom, instant clicks. 

The most common form of this is “Wishlists are inflated, velocity is the true indicator of success now”

Because of my survey and your generosity in sharing your game’s sales history, I have data from 160 games and I can actually test this hypothesis. 

Part 1: Is Wishlist inflation real?

Wishlist inflation is the theory that games these days are earning more and more wishlists because more and more developers are “wishlist maxxing.” (also known as doing marketing.) Maybe even some developers are using underhanded ways to get those wishlists.

The side effect of all these efforts (both honest and not so honest) is that wishlists are “worth less” so they convert worse. 

The survey has data spanning several years, so I can test how well a wishlist converted then vs. now and see if there is “inflation.”

I broke the period in half: 2020-2022 is the first modern wishlist maxxing era. The early pandemic era is when virtual festivals became a thing. The birth of Steam Next Fest and 3rd Party virtual festivals showed developers that they could earn thousands of wishlists in a week. A few enterprising developers knew about festivals, demos, and running sustained content creator campaigns.

The second era: 2023-2025 is when festivals really matured so far that Valve started to take a more selective approach to festival featuring on the front page of Steam. I didn’t include 2026 data in this analysis because we are still living through it and I can’t provide as many months of sales to compare. 

Comparing the two eras, you can see that it is true. There is wishlist inflation. 

Stats and Sample size

In the survey I have 160 games that were released in the 2020-2025 period (34 in 2020–22 and 126 in 2023–25).

So how well do wishlists convert? You can see the difference in the chart below. In the first month of release, a game in the 2023-2025 era sold about half (give or take because the range is wide) as many units per wishlist at launch as a game in the 2020-2022 era.

There is definitely wishlist inflation. It is real.

Theories?

I think there might be a few causes:

Valve made wishlisting easier

Valve worked with developers to make it easier to get wishlists. This chart comes from a 2021 Steamworks blog post. During the early era, Valve was touting how many wishlists they were pushing to games. They too were wishlist maxxing. 

For instance, they started adding little stars on hover tooltips (see screenshot below) so you could wishlist a game without even going to the Steam page. The wishlist became more of a “bookmark.” Adding a game to a wishlist used to be more cumbersome so people only did it when they truly liked a game.  

Democratized visibility via festivals 

In a previous blog post I looked at how many festivals Valve has run since they introduced the “make an event” tooling back in 2020. As you can see in the following chart, the number of games getting featured has exploded. 

We have suspected for a while that festival wishlists don’t convert as well. So more games have been getting more wishlists that don’t convert very well.

More developers are learning how important it is to get wishlists

There is better marketing information out there now, and more and more people know the importance of wishlists. So they are working hard to get them. There is nothing wrong with trying.

People know about Steam Next Fest. If you look at the rough number of games releasing every year and compare it to the number of games appearing in Steam Next Fest, you would see that for many years developers just skipped Steam Next Fest for some reason (probably naivety). With better information now, people are learning to do the basics like opt in to free wishlists earned from this event. This will lead to more wishlists with worse conversion.

Part 2: So Wishlist don’t matter anymore, right?

Again, this is where you can farm free engagement on YouTube. Just post a video of you scowling with your arms crossed with a label: “Wishlists don’t matter anymore, here is one weird trick.” 

TL;DR – nope, wrong. Wishlists are still key to predicting how well a game will do at launch. 

Let’s look at the real data. 

We can use correlation between two numbers to measure predictive power.

So first let’s determine if the first part is true. Is it true that “wishlists don’t matter anymore?”

This chart is a statistical analysis of how much the number of wishlists at launch correlates to units sold in the first month. A value of 1.0 for Spearman correlation means if we made 2 columns that ranked the top wishlisted games and then ranked games that sold the most units, the order of the two lists would be identical. R-squared means how closely games follow the trend line (a R2 of 1.0 would mean every dot is on the line).

Here is a scatter chart of the games analyzed, with a “best fit” line for each era. 

I like the lines because they make clear what the table can’t. Each line represents a different era. The lines are nearly parallel, so wishlists scale with sales the same way in both eras. The later line sits lower, which is the inflation I pointed out earlier.

Wishlists aren’t perfect, but they do predict how many units a game will sell.

But what about velocity

If you want to sound smart, you can post on Reddit “actually I think we are moving into a world where wishlist velocity is a more important indicator instead of raw wishlist counts.”

In this case I define velocity as the number of wishlists earned in the 30 days before the game launches. A game with “high velocity” gets a ton of wishlists right before launch. A game with low velocity gets a few.

Wishlist–sales fit (R², log-log) by launch era

As you can see, in the more recent era, velocity is not any better than total wishlists in correlation to unit sales. (Side note, yes velocity is 0.07 points lower in 2023-2025 but difference is statistical noise, considering there were far fewer games in the survey for the 2020-2022 era.)

BUT, if we look at all games smooshed together across all eras, we can improve the units sold prediction a bit more if we look at both total wishlists AND velocity. 

Wishlists are still king

It is hard to wrap your mind around it… but both things are true: wishlist counts have inflated (each wishlist sells fewer units), but wishlists predict sales just as well as they did 3 years ago (the 0.05 difference is within the noise for this sample size.)

As an analogy to the real world… just because the cost of living is going up doesn’t mean we don’t have to pay rent.

Unfortunately, we still need to count, collect, and talk about wishlists. Sorry, you can’t make any cool viral YouTube videos about wishlists not mattering. Life is boring. 

Prove me wrong

I am still trying to collect more data from more games. So if you have launched a Steam game in the last 6 years, please submit your data. 

Official HTMAG Steam wishlist and sales survey