Thread by @Enichan

So some people have asked me for a good summary of what's going on with Steam just reaming indies unless they're in the 1% of popularity.

Here's my thread on it.

I'll start with a disclaimer, observations, and then theories floated by devs including myself.

DISCLAIMER: Valve doesn't communicate about their algorithm changes, and devs don't talk about their experiences publicly as a rule. This information has been collated from talks with devs within my whisper and back channel networks, as well as my own experiences.

Observations:
In October something happened to Valve's algorithm that caused organic traffic (and revenue) to plummet overnight, for games that had been doing just peachy before, for a long period of time. I've heard accounts between 33 and 80%.

I was not personally affected.

Valve did eventually admit to some kind of bug in a change they'd made, and later said they'd resolved that bug. For some devs traffic returned (somewhat). For many (most?) it didn't.

Affected devs noticed total outstanding wishlists dropping for the first time EVER.

Their organic traffic had been impaired strongly enough that more wishlists were being deleted and converted into sales than new traffic resulted in new wishlists.

Before this I'd never heard of a game where this happened. Either it stays steady, or it goes up, never down.

Somewhere between October and December more people started being affected by these algorithm changes, me included. In December we ran a week long sale, and the winter sale happened, and our numbers were dreadful. Both together earned less than a week long sale had months before.

The winter and summer sales are normally the two times a year when devs make THE MOST money on Steam in their longtail. This year it was extremely slow.

Customers also started complaining on Twitter about bad discovery queue/recommendations, less indies they wanted, more big AAA

Somewhere in that timeframe (I can't remember the actual time) devs communicated with Valve but Valve said everything was working as intended and they would do nothing further, opting to wait instead for their machine learning algorithm which will supposedly happen some day.

Many devs hoped that the winter sale performance was a fluke. The Lunar sale dispelled most of us of that illusion. Some devs said that sales during the Steam sale were virtually indistinguishable from a good non-sale week before these changes happened.

Personally I held out hope until I started running another week long sale, this week. We even did a patch with a seasonal event to bump interest and player numbers, and through some fluke had a few reviews come in right before.

It didn't help much.

We ran week long sales in October in 2017 and 2018, lets compare their performance to the winter sales shortly after (numbers are gross):

2017:
Oct: $2,954
Winter: $4,632 (+57%)

2018:
Oct: $1,294
Winter: $1,682 (+30%)

This might not seem like a big drop, but this is a big deal

Devs will reinvest in new products based on historic longtail revenue trends. If overnight that expected revenue gets cut by 30% or more, suddenly devs can find themselves in some very hot water.

Who is this happening to? Devs across the spectrum. Devs who've spent years on a previously thriving early access game with sudden reduced growth, devs who's been in this business for over a decade full time, devs who were, almost universally, doing just fine before October.

Theories:
Valve changed the algorithm, we know that much, but how? Many of us suspect that pressure from Epic (see the AAA revshare drop shortly before the October thing) and decided to give less organic traffic to smaller/less popular titles, and more to popular AAA titles.

This is consistent with the issues users were having with their discovery queues back in December. Some of the changes to discovery queues may have been rolled back (based on anecdotal data) but complaints remain, so definitely not all.

Speculation is that someone looked at where they get the most money, and decided to focus on that, to the exclusion of indies which draw many consumers to the platform and have a high but not directly monetary value.

Given my failed experiment the change is probably not based on a game being recently updated. There's some speculation that active player counts factor into it, and my personal pet theory is that 250 reviews is the new 50 reviews.

(Aside: there is what is called the 50/500 rule, that games receive large jumps in visibility at 50 reviews and then again at 500 reviews, Valve denies the existence of any such metric in the algorithm)

Player counts may be part of it even though my experiment increased player counts but not visibility. Maybe there's some metric that smooths out player count spikes (during sales) or maybe I just didn't manage to hit whatever lower bound is required.

One of the more interesting theories is that the October bug 'polluted' the algorithm's historical data somehow, and Valve literally can't do anything about it without wiping all algorithmic data to a clean slate, which they can't do. And if you have "the bug", too bad.

Whatever is going on, many devs are on "life support"; getting far more of their sales from previous wishlists than before, because their organic traffic has been cut. This isn't sustainable.

In 2018, 80% of our revenue from sales came from Steam. 7% came from @itchio, 5.5% came from humble, and 7.5% came from @chronodeals. Excluding Chrono, 86% of our sales revenue came from Steam.

A 30% reduction in Steam revenue is a 25% reduction in overall revenue.

The vast majority of devs can't have their revenue cut by a quarter, third, half, or even two thirds overnight and survive for very long. Margins in this business are already marginal, and survival is already an extremely difficult task.

My prediction is that a lot of indie devs, some of which might be favorites of those reading this, are hanging on hoping for a turn-around before the summer sales, or for Epic to open up their store to them.

I expect a lot will start going out of business starting July.

This trend is extremely concerning, because of the economic fallout for many indies in our industry, lost revenue from supporting service providers, but also because it shows how fragile our existence is and how much we're at the mercy of Valve's capricious whims.

I've been ringing the alarm bell since December. Other devs who've been talking about this are @glassbottommeg, @GreyAlien, and to some extent @unormal.

If you're press, and you want to talk to me about it, please email me at [e-mail] or DM me, my DM's are open.

Please keep in mind as you read this that most devs will NOT talk about their financial difficulties in public, believing it makes them look weak, hurts business opportunities, and reduces sales. They may not be wrong about that, but this is important, and so I'll take that hit.

PS: I tweeted so fast by using the little + button when you compose a tweet that lets you compose a tweet thread all in one go, at least on the web. I dunno if mobile clients have it.

Now you know.

I guess this solidly debunks my "250 reviews is the new 50 reviews" theory

https://twitter.com/GreyAlien/status/1108817662153551872

Post-disclaimers:
People will say "those games are bad", but these games were fine one day and not the next, so that's nonsense. Others will say that revenue was going down anyway, and it has been ever since Greenlight/Direct, but it collapsed overnight, so this wasn't organic.

This has nothing to do with a failure on the side of devs, cause these were devs with games who were already succeeding, and had a history of years of being a success. This wasn't a slow collapse, which only leaves the possibility that Valve DIRECTLY did this, and doesn't care.

In the short term? Valve can fix what they broke, but they won't. Epic can open their store doors, but they won't until much later this year.

So... extreme public pressure on Valve? A concerted effort of devs, consumers, and press to shame them for this?

https://twitter.com/feureau/status/1108818660787777536

They should be shamed for this, they've instigated an extinction level event for indies and then just shrugged their shoulders and went "working as intended".

Long term? Competition from Epic and a broader consumer awareness of these issues leading to growth of other platforms.

I'd like to reassure everyone that despite this downturn Kitsune Games is going to be okay. We've planned for this since Steam Direct.

However it's still a blow, so if you're very concerned now is an excellent time to kick us a buck or two on Patreon patreon.com/sharkhugseniko

This is interesting, this shows the green "other product pages" line (which I believe is links between similar games?) flatlining like has happened to many afflicted games. But this apparently started THIS MONTH.

I... I have no explanation for this.

https://twitter.com/SuperSixStudios/status/1108943172464312320

Reminder that this change that is destroying pc revenue for all but the lucky and most successful indies is "working as intended" according to Valve

https://twitter.com/tomerbarkan/status/1109064734454751232?s=19

Working as intended according to Valve, we weren't selling well enough so they took away all our visibility

https://twitter.com/nyphur/status/1109138969877467136?s=19

All Valve apparently really has to do to fix this is to give a boost to tags and related game visibility. But they won't because they wanna wait for their machine learning algorithm.

When will this be done? Nobody knows but at this rate when that comes out it'll be too late.

Please note I don't want Epic to kill Steam. Then we're in the same spot all over again but at Tim Sweeney's mercy. I want Steam to stop sucking so there's healthy competition because it benefits devs and consumers.

Although Steam does need to ditch the 30% thing, like come on.