The long tail of casual games

[Thorbrian]

Hey all,
I've read various posts here and there about how well games sell over their lifetime ("casual games have a long tail", "portals only care about new games", etc) so I figured I'd throw some data into the ether on that topic.

The attached image is a normalized graph of the average sales per week of a single game over it's lifetime in the Reflexive Arcade system (we have about 4 years of usable data for this purpose). I got it by grouping all the sales by the age of the game in weeks, then dividing by the number of games that were that old. Then I normalized the values so that 1 represents the sales rate that the graph seemed to flatten out at. I didn't do anything to try and account for and remove growth of the affiliate network or hit games. So the graph incorporates a number of factors, and it gets noisy near the 4 year mark because we have many less games to average over.

So for us, games have sold about 10x when first released, but quickly dropped down to a stable rate, which it seems like they could sell at forever. My expectation is that other portals and direct sales systems would show similar behavior, but I am curious if anybody else has had a significantly different experience.

[AJirenius]

I guess it must be a matter of how a portal decides to make older games available. At the big portals I guess all advertising and space goes to the new games and a 2 year old game have extremely less exposure than it would have in a smaller portal with fewer games where people actually can have the energy to browse them all and find this "jewel among games".

I guess the accessibility to the older games is very critical when it comes to a lifelength of a game. I'm sure you could take almost ANY old games (say 3 years old) and republish it and it would hit some great sales again.

[berserker]

I doubt 3 year old game will be THAT great seller as quality bar have risen since then significantly however some portals like AOL indeed feature quite old games from time to time and even rejecting some sequels due to originals selling quite well to their taste.

[Nexic]

Reflexive and ArcadeTown both seem to have good long term sales, with others it's all or nothing in the first week. At least that's my experience.

[woo]

Great chart. That definitely supports the idea of getting to the point of having multiple games on the market in order to be successful.
Yeah, I know - "Easier said than done." :)
-Andrew Douglas
http://theoreticalgames.com

[Davaris]

When people ask if they should quit their day jobs to make their dream game, we should point them to this chart.

There's nothing like a healthy dose of reality. :)

[soniCron]

Great chart. That definitely supports the idea of getting to the point of having multiple games on the market in order to be successful. I'm not sure how you came to that conclusion?

[EnigmaCEO]

When people ask if they should quit their day jobs to make their dream game, we should point them to this chart.

There's nothing like a healthy dose of reality. :)

I wouldnt go that far. Just shows that depending on portal sales of one title might not be the best financial move. I just recently quit my day job and I sell well over 10 games a day everyday. My goal has been to make 1 game every month (or two) and so far its panning out. Sales haven't dropped off on my oldest game even after a year on the market.

My PC ventures haven't been so great yet but I'm working on it :)

[woo]

Okay. Success isn't necessarily due to one big hit (that tops the charts for it's first month). The graph certainly doesn't indicate success in terms of total sales - for example if I actually only sold 25 copies or so in the first month, this graph would indicate that I could expect sales on the order of 4 copies each month after the first year. That's not a lot of sales no matter how you slice it. But!
The long tail effect means that having lots of "smaller" but essentially infinitely recurring revenue streams can quickly add up to more revenue than, say, even your newly released game. So let's say you have, over the course of 3 or 4 years, as covered by this graph, 5 titles that were of similar quality and with similar initial sales - they each sold (for the sake of making the numbers easy) about 100 copies in their first month (still not a huge number mind you) and about, as the graph would indicate, on average, 15-20 copies per month every month after their first year. That means that after 3 or 4 years you're actually selling 100 copies a month for your existing titles even if all of them are 6 months old or more.
So is it possible to be making next to nothing even with 5 projects on the market? Sure, but due to "economies of scale", it's costing you essentially nothing for those games you produced a year ago or more, and if you have enough of them, they can be bringing in revenue that matches the sales totals of even your best month of your newly released game that you spent X months slaving over.
These numbers are certainly skewed towards a particular market and other portals certainly may perform differently, but let's say that graph showed something like a 10x drop off between your first month and your recurring monthly sales, or, if after a year or two the number dropped to next to 0, that would seem to indicate that success is based on the "what have you done for me lately" business model which is a lot harder to make a living off of.
-Andrew

[nights]

Great post Thorbrian!

However, the 90% drop in the first year sounds a little bit dramatic. Isnt it more like 30% off annually for say 3 years...?

[James C. Smith]

This chart sure makes the first 5 weeks look very important, and they are, but it also makes the 3 years look important to me. I am too lazy to do the math, but I would love to know what percent of the game’s sales happen in the first 5 weeks vs. the rest of the first 3 years. Or the first 3 months vs. the rest of the first 2 years. It seems to me that those huge sales during the first 5 weeks will, in the end, only be a small percentage of the games overall sales.

[Thorbrian]

Here is another perspective on the same data, it's the cumulative total of the sales:
484
To answer James' question, it looks like 10% of the sales over 4 years came from the first 5 weeks. If any one wants to play with some other data points, here's the source data for the graphs (tab-delimeted text):
485
Note that the data is not "raw" data - it's the normalized data (where 1 represents the average steady rate of sales after the first year).

[Thorbrian]

When people ask if they should quit their day jobs to make their dream game, we should point them to this chart. I don't see how these charts address that issue at all. It's just one data point on how games sell over time and it's normalized to remove scale. It makes no statement on the total profitability of making games except at how that total profitability may come in over time. There are likely other charts and datapoints that would serve that end very well, though.
However, the 90% drop in the first year sounds a little bit dramatic. Isnt it more like 30% off annually for say 3 years...?sorry, I should have been more specific. I just meant that the weekly rate in the first week was, on average, nearly 10x what the (presumably) stable sales rate was. I think what you are talking about is how the first years sales compare to the total sales over four years? If it is, then it's about 50% in the first year - or to put another way, on average the first year has sold 3x the copies of each year following.

[Coyote]

Thorbrian - that is an outstanding resource, thank you for sharing it!

Both of those images together tell a very interesting tale (tail? :) ) --- at around week 61, the sales had dwindled off to the "stable point" - a point where, in conventional brick & mortar sales, the game would be dropped off the shelves (or dumped into the discount section). Yet at that point it has only generated a little more than half of its total sales in one third of its life-cycle (thus far).

Being successful as an indie really does seem to mean being in it for the long haul.

[cliffski]

Thats because you have your own shelfspace at your own website. Granted, this just shows portal sales, but I know that as far as positech.co.uk, it will be at least another year before Democracy is taken off the front page, and at least another 2 years before kudos is taken off.
And thats an optimistic case, assuming I write 2 or 3 more games in that time that outsell both titles.
I don't think indie games age that badly. I fully expect Democracy and Kudos to sell the same next february as they do now (from my site). I think at the 5 year point stuff shows it's age, mainly due to screen resolutions going up, and new operating systems coming out, but apart from that, unless you have hit saturation point exposure-wise, what's to stop a game selling from your own site?

[arcadetown]

It also depends a lot on the title, the portal/site, and it's ability to maintain sales and earn promotional space there. For example, World Domination and Age of Castles still do very nicely here even years later. Venturing an educated guess I'd say the top 5% of games continue doing very nicely for a long time, top 25% stabilize around 25 - 50% of launch sales, and rest tend to drop off pretty quickly.

[Davaris]

don't see how these charts address that issue at all.

Oh I see! Sorry I didn't understand it. I thought those were average sales per day, which would be very ordinary. As you can tell, statistics isn't my subject.

[woo]

While I certainly believe that arcadetown is accurate, especially at certain portals, it would be very interesting indeed to see an "overlay" of those figures for the top 5% and the top 25% of games in relation to the same "normalized" baseline. A line for "everything below 25%" would be interesting as well. If that makes any sense. If the top 5% for example are really carrying the weight in order to keep the "average" up, then that would definitely impact the interpretation of the chart.
-Andrew

[Pyabo]

it will be at least another year before Democracy is taken off the front page

You should just re-release it every 4 years with a new ordinal...


(yes, yes... you're British... I know)

[Thorbrian]

I did something along the lines of what woo asked for - I broke stuff down by top 50 sellers and all below the top 50. Also I broke stuff down by reflexive developed vs. all other developers. All lines are normalized (which is why the all games line can be higher than all the others sometimes - before normalization the All games line is always between 2 lines that partition that space)

494

[woo]

Thanks! That's definitely less separation than I was expecting for sure (heh.. it looks like maybe there's a bit of preference given to Reflexive games for the first year maybe? :) ).

The drop off at 3 years or so was a little odd though - it looked like everything dropped in relation to the base line, so I'm not sure how to interpret that. Any ideas?

Thanks again. It's really great that you're willing to share this info with the community!

Now, about those actual sales figures .... :)
-Andrew Douglas
http://theoreticalgames.com

[Thorbrian]

it looks like maybe there's a bit of preference given to Reflexive games for the first year maybe?There is definitely a Reflexive-developed game bias to our affiliate systems first year's performance, but I think it's a result of not having many affiliates and not having many games (aside from our own) when we first launched the arcade & white-label affiliate system.

The drop off at 3 years or so was a little odd though - it looked like everything dropped in relation to the base line, so I'm not sure how to interpret that. Any ideas?well it was right about 1 year in to the life of the arcade (1.5 months shy of 3 years ago) that we released Ricochet: Lost Worlds in our arcade system... The sales from it were much larger than our expectations, and the release correlates with a significant growth period for the system in all respects within a couple months following. I would expect that is a major contributor to the data coming out the way it does.

[terin]

I'll add a note to this thread-

There are certain titles that are very hard to qualify with a statement like "top 10" or "top 50" that sell very well for long, LONG periods.

So while the average you may get from a portal is 1 sale a week, for instance, it is my experience that this is true for about half the games. There is another half which just vanish into almost complete obscurity, lets say 1 sale a month or less. The other half, which hopefully you are a part of - lets say 1-2/week for years and years.

JUST like the top 10 syndrome there is a SMALL % of games that just do really well for long periods. Lets say it sells ok in the first 5 weeks - probably top 50 somewhere, probably top 30 actually. It will then go on to have stable sales for YEARS. I would say 10x or more than the "normal" game (so in this example 10-20/week for years - but whatever, 10x+ from what most games on that particular portal do). This is rare even for the top 50, maybe a handful of titles a year do this.

So its quite possible to create a game on a portal, MISS the top 10, and still go on to outsell a top 10 game by a huge margin. I've seen a lot of games follow the pattern of big sales, a fairly quick slide, and then stable low numbers for a long time... but occasionally we catch one that has a bunch of sales, a short slide, and remains in the top 30 or so sellers for ages and ages.

A noteable pair, as Brian mentioned, is World Domination and Age of Castles - which are pretty consistently in the top 20 games every day ; but never made that huge splash of sales that everyone claims is so vital to success.

I'd say this is more rare than a top 10 game - but I would be willing to bet in the long term THIS is the best way to make money and should be the real goal of every developer. An evergreen title.

Though I suppose the very best would be top 10 AND an evergreen... looks like Virtual Villagers and MCF: Huntsville are likely examples... time will tell though; for now both titles have maintained top 10 since launch with us.