Why I trust Valve's judgment on Seasonal Sales

Last year, Valve announced they would no longer do Flash sales (ie, short-lived mega discounts on selected titles during a sale itself) during their popular seasonal sales.

Loud voices on the internet were unhappy, the top voted comment on this reddit/r/games thread was:

Well, the sales have come and gone, and the results are in:

This image in particular tells the story:

But why take Valve's word for it when you can take mine?

Valve doesn't let us share sales stats directly, but they have given us written permission to share traffic stat reports, a feature they enabled with the Discovery Update.

Spoiler alert: sales are highly correlated with store page traffic.

This is particularly relevant because as we mentioned in a big blog post last month, our game is over four years old and still selling a decent number of copies. For us at least, the Stegosaurus Tail lives on.

To be sure, our sales curve has been feeling a little limp lately, which is why it's high time we got our next product out the door. But I'm still flabbergasted that we've been able to last this long, and without the Discovery Update and especially the recent tweaks to the seasonal sales, we'd be in a very different place.

Valve has gotten a lot of criticism as they've gone from scrappy underdog to hulking overlord, and there's definitely a million ways in which they could and should improve, but one facet of Steam that hasn't been neglected is this constant drip of improvements to the store's discovery engine.

"Discoverability" is not an easy problem to solve -- the more titles in your library, the harder it gets, and exponentially so.

Curated stores like GOG solve this issue by being very selective about who they let in -- and for those of us fortunate enough to get in, GOG is indeed excellent. But Valve are trying to walk that fine line of being open to basically everyone without drowning everybody in obscurity thereby.

It's a "damned if you do, damned if you don't" situation -- it's popular now to complain about how crowded Steam is, but do you remember back in 2012 when everyone was outraged at how critically acclaimed Indie titles weren't making it onto Steam? We can complain all we want about Greenlight and Valve "opening the floodgates", but if they hadn't, beloved mega-successful titles that came out of nowhere like Undertale and Five Nights at Freddy's could have easily been stuck in limbo.

In typical Valve fashion, the solution is to throw something at the wall, see if it sticks, and if it's a huge mess don't worry too much and  just learn from it and carry on. This is in sharp contrast to say, the console industry which is now playing catch-up with the PC market.

Valve could have just done things the traditional way - give more attention to the games that are already selling a lot of copies. That's what the app store does, after all. Instead they constantly mix up their formula, even when all the loudest voices on the Internet say they're wrong.

Check it out:

This particluar feature is new for the Lunar New Year sale. (Wikipedia tells me Lunar New Year is a major holiday in Korea and China, and a minor holiday in Japan).

Our game happens to be localized into both Korean and Japanese (sadly not Chinese), so we'll be very interested to see if we see a sudden spike in sales from those regions (we've seen steady organic growth in those regions already just since we added the game localizations and store descriptions).

Things like this seem basic to the point of obvious, but every little feature adds up, and most stores don't even do them!

It just makes so much sense that you should take basic information about a player into account when deciding what games to show them on the front page, rather than just blasting everyone with Call of Duty and Civilization V for years on end. Here's a free tip for budding game stores: if I have a title in my library already then stop advertising it to me!

Let's go back to that traffic report from before. Here it is without the editorial overlay:

We don't have any traffic reports from before the Discovery Update, but I did some revenue comparisons that detected a significant bump -- I covered this in I Wish Upon A Steam.

The giant spike in Halloween 2014 is a bit of an anomaly -- Steam had a very simple structure for that sale, basically just a hand-picked list of games with "scary" themes, and it was quite short, so we lucked out by getting tons of traffic. The 2015 Halloween sale was similar but more inclusive so we didn't get as disproportionate a share of the traffic. The orange you see in the Halloween sales is "search" traffic, but that's misleading -- the front-page lists that Steam generated for the Halloween sales simply piggy-packed off the search engine (it was just a sorted search list with pre-applied Halloween tags). So really that's traffic from "you were picked to be on a prominent list," not "users organically found you via search."

Throughout the year no event besides sales promotions really moved the needle except for this blog post we wrote which garnered a surprising (but comparably modest) amount of traffic.You can also see quite clearly that traffic doesn't seem to be noticeably greater or lower after a sale -- there might be a slight dip the next day, but not much more than that.

2015's fall sale was quite similar in structure to 2014's, and they performed about as well, although 2015's lacked daily deals or flash sales. Holiday sale was the big difference - not only were the daily deals and flash sales gone, but rather than a minigame or other gimmick to "increase engagement", Steam added a simple trading card incentive -- go through your discovery queue, get a trading card. Limit X per day. Come back tomorrow and do it again. Simple, and effective, and the numbers speak for themselves. We saw a huge spike in traffic, nearly all of it from the discovery queue. And yes, that traffic directly drove a roughly comparable amount of sales.

I'm really glad to see Steam moving in this direction. Not only were flash sales and daily deals during a seasonal sale bad for customers (you had to constantly watch the sale and be afraid of missing out), they were bad for developers. It created a two-tiered system where everybody was on sale, but only a select few were really on sale, and it seemed like the same list of big name developers would get the special flash and daily deal slots all the time. Even worse, it created a hostile mentality where developers felt like they had to aggressively compete on discounts in the hopes that they'd be picked for the spotlight. Now it's much better -- sure, XCOM 2 will probably get a big hand-picked headline, but if you're Korean and like Tower Defense and RPG's? You just might see Defender's Quest. The games you see on your front page now depend mostly on you.

And that's as it should be.