Archive for the ‘Game Development’ Category

The cult of Steam wishlists

Monday, April 6th, 2020

Been meaning to talk about the “cargo cult” of Steam wishlists for a while.

Wishlists were originally a genuine expression of interest in purchasing a game. A couple of years ago I worked out a formula that said:

pre-launch wishlists x0.5 = week1 unit sales

Since then there’s been a lot of talk about the importance of wishlists by devs and publishers. And so devs have been trying to get as many wishlists as possible by any means possible.

Ancient Enemy pre-launch lifetime wishlist actions

Quality vs. quantity

But what is not often discussed is wishlist QUALITY.

In advertising it’s well known that some clicks are better “quality” that others. Some click sources end up converting into sales at a much higher rate than other sources.

My theory is that wishlists are very much the same in that they have a “quality”.

Some wishlists are from true fans (especially if you’ve shipped games before) and others are from people who aren’t really interested and will wait until the game is 90% off or may never buy it.

Also these days devs beg for people to wishlist their game as a favour.

And gamers think they are helping you out by clicking “wishlist” even though they may have no actual intention of buying it.

So devs blast news about their game everywhere and may end up getting an inflated number of low quality wishlists.

How does conversion work?

This makes the dev feel reassured especially if they’ve seen my formula. However, those wishlists are not an accurate representation of interest any more and the dev may be in for a shock when they launch as the conversion could be way lower than 1 sale per 2 wishlists.

People often misunderstand how my formula works so please let me clarify:

Your pre-launch wishlists do NOT convert at 50%. What happens is they convert at about 10% (on average) and then you get 4x that amount of organic sales. This = 0.5 sales per wishlist.

That’s because the wishlists indicate interest in your game but you make most of your sales organically when it hits the store.

But if you have super-inflated your wishlist count with low quality ones, not many of them will convert and your organic sales will be much lower too.

Other factors

There are other factors here such as if your launch day offering matches expectation in terms of price (which is often unknown at the time of wishlisting), and maybe what else is launching that day/week, etc.

BUT it’s not totally fruitless to try to get loads of wishlists because there is some evidence to suggest that the Valve algorithm will boost visibility for “popular” games with a lot of wishlists, so that’s worth aiming for, although there are no hard numbers to aim for.

Also you will have increased the visibility of your game and so potential customers will be more familiar with the logo/screenshots and may purchase it at a later date. Plus you’ll have practiced marketing (found out what worked/didn’t) and refined your message.

It’s still probably worth trying to get lots of wishlists but if you are very successful at this bear in mind that they may not all be good “quality” and may not convert into that many sales.

Good luck!

Ancient Enemy – complete!

Tuesday, March 17th, 2020

BIG NEWS: Ancient Enemy is DONE! It’s currently being beta tested by a bunch of lovely people.

We will be releasing a trailer and announcing a launch date soon.

Meanwhile please spread the word so we can get more wishlists, this makes a big difference to our visibility on Steam. Thank you!

Steam Sale Wishlist Conversion Rate Analysis

Monday, October 7th, 2019

Previously I’ve written about the importance of collecting Steam wishlists before launch and about how well those wishlists may translate into sales at launch, but in this post I wanted to share my data about wishlist conversion rates during discount sales AFTER launch.

Collecting Data

In early 2016 Steam started reporting how many wishlist emails were sent out when you run a discount (e.g. during a weeklong sale or a Steam sale) and how well those wishlists converted into sales.

You can find the data for your game in the sales reporting system (not Steamworks). Click on your game and then click “view detailed wishlist breakdown”. Choose a date range of “all history” and then scroll to the bottom of the page and you should see a table like this: 

I’ve used the 7-day conversion rate for the charts in this article. I also included the discounts I used on a couple of charts. You can find those in Steamworks by clicking your game and then clicking the well-hidden “add or edit discounts” button next to the pricing button.


Here’s the chart for Shadowhand. It launched with a 15% discount and then we stepped up the discount rate quite quickly to 50%.

Those early sales did pretty well but then the wishlist conversion rate dropped after the famous “October 2018 algorithm bug” (data point 7). I can’t say for certain it had an effect but it sure looks like something happened. Anyway, since then we’ve gradually pushed up the discount but have only managed to stabilise the wishlist conversion rate at about 0.9% rather than see any gains.

The overall conversion rate is 1.22%. So, for every 1000 wishlists the game has we make 12 sales, woo!

It should be noted that normally in a sale you also make “organic” sales that do not come from wishlists. Those can vary quite a lot (per game and per discount) but as a ballpark I find that wishlists account for about 50% of overall sales during a sale. Perhaps that’s a topic for a future article…

Pre-launch vs Post-launch wishlists

We had about 4000 wishlists before launch and they shot up to about 13,000 during launch week and then continued to climb rapidly through the Steam Winter 2017 sale to about 21,000. Most games exhibit similar behaviour on launch but at different scales.

In 2018 additions slowed down but were still climbing until the Oct algo bug when they began to decline slowly due to decreased visibility on the store (see the purple Outstanding Wishlists line below). We weren’t the only game to notice this. 

Pre-launch wishlists can convert at about 10% on average, sometimes higher, sometimes lower. But after launch they won’t convert anywhere near as high as that; more like 1-2% on average. Post-launch wishlists can convert higher at around 3%-4% in your first few sales after launch, or with a steep discount later on, but those events are rare and the average is likely to be a lot lower. I say this having studied the data for 4 of my games over several years (more charts below).

Regency Solitaire

Regency Solitaire launched on Steam in May 2015 which is before the wishlist conversion data was available, so this chart is missing the first 9 months of sales. The average on this chart is 1.63%.

Spooky Bonus

This chart is also missing the first 4 months of data. I haven’t added in the discount percentages. The average is 1.19%

Spring Bonus

This data includes the first sale after launch which converted surprisingly well (5.6%) and during a later high discount but this game doesn’t have very many wishlists so the actual revenue was low. The average is 2.1%


For the four games above the average wishlist conversion rates range from 1.19% to 2.1% but it should be noted that they start higher than that and drop over time. The worst conversion rate I ever had in a sale is 0.2% and the best is 5.9%.

Don’t forget that just because you have a better conversion rate and in theory sell more units, if your discount was high, your revenue may not actually be any better than in previous lower discount sales.

Also, once you use a high discount you can’t really go much lower again (I confirmed this via experiments) otherwise your conversion rates will suffer. Furthermore, higher discounts lead to worse reviews on your store page because those customers are less invested in their purchase, so be careful.

Anyway, I hope you found this post interesting and can use the data to give you some idea about how your own games may perform. Also, if you have any of your own data to share, please do so in the comments.