Reading Steam wishlist data by country and turning it into decisions
Somewhere in Steamworks there is a breakdown of your wishlists by region, and the first time you look at it the reaction is usually the same: interesting, but what am I supposed to do with this? The countries at the top are the countries you would have guessed, the long tail is full of numbers too small to mean anything, and there is no obvious line between what the report says and what you should build next.
The reason it feels unusable is that the report is evidence, not a verdict. On its own, a country ranking mostly reflects where Steam users are — which is the same for almost every game on the platform, and therefore tells you nothing specific about yours. The information is in the comparisons: your share against the platform's share, one region's behaviour against another's, and the same region before and after you changed something.
This article covers where the numbers come from, why country is not the same thing as language, how to read the breakdown against a baseline, and how to get from a table of regions to a defensible decision about which language to add next. The worked example partway through uses invented numbers, and is labelled as such — the method is the point, not the figures.
Where the numbers live and what each one measures
Steamworks reports several different things that all get loosely called wishlist data, and they answer different questions. Report names and layouts change over time, so rather than memorising a menu path, learn what each kind of number represents and find it under your app's sales and traffic reporting.
Store traffic tells you how many people saw or visited your page, usually with a breakdown by region and by where the visit came from. This is discovery: it measures how well Steam and your marketing are putting the game in front of people.
Wishlist activity separates additions from deletions and from purchases or activations. Additions are the interesting signal for this exercise, because an addition is a deliberate act by someone who read the page and decided the game was worth remembering. Deletions matter too — a region that adds and then removes at an unusual rate is worth a second look, because something is changing people's minds after the fact.
Sales by region come later and are the ground truth, but they arrive too late to inform a pre-launch language decision and they are entangled with regional pricing. For deciding what to translate, wishlist additions relative to page visits is usually the most useful pairing you have.
- Visits or impressions — did anyone see the game at all in this region
- Wishlist additions — did the people who saw it act on it
- Wishlist deletions — did they change their minds afterwards
- Purchases and activations — the eventual outcome, useful after launch
Country is not language, and the difference matters
The report gives you regions. The decision you are trying to make is about languages. Those two are related but not interchangeable, and translating one into the other carelessly is the most common way this analysis goes wrong.
Some languages are spread across many countries. Spanish-speaking wishlists arrive from a long list of separate rows, each individually small, which makes the language look weaker than it is if you only ever read the table one line at a time. The same happens with Portuguese, Arabic, and French. Conversely, one country can contain several languages you would translate differently, and Chinese is the case that catches people out most often, since the choice between Simplified and Traditional is not a single decision keyed to one row of the table.
There is also the plain fact that where someone is does not determine what language their Steam client is set to. Players living outside their language's home region, bilingual players who keep their interface in English, and regions with large diaspora populations all blur the mapping in ways no report will disentangle for you.
The practical fix is to aggregate before you interpret. Group the region rows into language buckets first, using your own judgement about which language each region most plausibly reads, and then read the totals. And for a global baseline of what the platform's language distribution actually looks like, the Steam hardware and software survey publishes client-language shares — that is language data rather than country data, which makes it a closer proxy for the question you are asking.
Compare against a baseline, not against zero
Here is the trap that makes raw regional tables so misleading: almost every game's country breakdown looks broadly similar, because it is mostly a picture of where Steam users live rather than a picture of who wants your game. If you rank your regions by wishlist count and act on the top of that list, you are mostly acting on population.
The informative quantity is whether a region is over- or under-represented relative to what you would expect. If a region accounts for a certain share of the platform's users but a much smaller share of your wishlists, something specific to your game is happening there. If it accounts for a much larger share, something is working — a streamer, a community, a genre affinity — and it is worth understanding what.
Under-representation has several possible causes, and language is only one of them. Your genre may simply be less popular in that region. Your price may sit badly against local expectations. Your marketing may never have reached there. Or your page may be unreadable to them. What separates these is the ratio between visits and actions: if people are not arriving at all, that is discovery or genre; if they arrive in numbers and then do nothing, that is the page.
A worked example with invented numbers
The table below is fictional. The figures are made up purely to show how the comparison works, and they are not measured from any real game — do not treat them as benchmarks or as anything you should expect to see.
ILLUSTRATIVE EXAMPLE — invented numbers, not real data
Region Page visits Wishlist adds Adds per 100 visits
---------------------------------------------------------------
Region A 40,000 2,000 5.0
Region B 22,000 1,320 6.0
Region C 18,000 180 1.0
Region D 9,000 540 6.0
Region A: most visits, most wishlists. Your biggest market, and it
tells you nothing you did not already know.
Region C: the interesting row. People arrive in real numbers and
almost none of them act. That gap is a page problem, and
'the page is not in their language' is the first
hypothesis worth testing.
Region D: small but converting as well as your best region. A good
candidate for growth, because the page already works
there and only discovery is missing.Turning the reading into a decision
A row that converts poorly despite decent traffic is a hypothesis, not a conclusion. Before spending anything, check the alternative explanations, because acting on the wrong one costs real money and produces no change.
Ask whether a spike in that region came from a single source — one streamer, one aggregator post, one festival — because traffic from a single viral moment converts differently from organic store browsing and will distort any ratio you compute during that window. Ask whether a sale, a price change, or a platform-wide event overlapped your measurement period. Ask whether you did marketing in one language that would naturally skew the whole table.
Then choose the cheapest action that could actually change your mind. For a language question, that is almost always the store page rather than the game: it is a few hundred words, it sits upstream of every other number, and if the hypothesis was right you should see the visit-to-wishlist ratio for that region move without anything else about the game changing. If the hypothesis was wrong, you have spent very little to learn that the problem was discovery or genre rather than language.
- High visits, low wishlists — start with the store page in that language
- Low visits, healthy conversion — this is a marketing and discovery problem, not a translation one
- High wishlist deletions — look at what happens after the page: price, release date changes, or a demo that disappointed
- Everything small and flat — you probably do not have enough traffic yet for any regional read to be meaningful
Measuring whether the change worked
Once you translate the page, resist judging it on the first week. Store data is noisy at the scale most games operate at, and a single week can be dominated by something unrelated. Take a baseline from several quiet weeks before the change, then compare against several quiet weeks after it, and compare the ratio rather than the raw count so that a general traffic increase does not read as a success it was not.
Change one thing at a time if you can. Translating the page, adjusting regional pricing, and running a discount in the same week means you will never know which of the three did anything. Doing them in sequence takes longer and is the only version that produces knowledge you can reuse for the next language.
Accept that you will never get a clean experiment. There is no control group, seasonality is real, and the platform's own promotion changes underneath you constantly. What this means in practice is to look for effects large enough to survive that noise, and to be honest when a result is ambiguous rather than reading a story into a small movement.
Finally, write down what you did and what followed. A short record — the date you translated the page, the region, the before-and-after ratios, and your conclusion — turns a series of one-off experiments into an actual understanding of your own game's markets. That record is what makes the third language decision faster and better founded than the first one.