Author: Jarek Jachna
On April 7, Steam officially announced an update to its platform and the expansion of Steamworks with new analytics tools to measure the effectiveness of marketing efforts. This is news that both we and our game development clients have been waiting a long time for.
How do the introduced functionalities influence the effectiveness of campaigns aimed at building a wishlist? In this article I describe the new features and share the results and observations from a campaign conducted for our client. The first one during which we used Steam UTM Analytics.
What is Steam UTM Analytics?
Steam UTM Analytics works on the basis of the mechanism on which Google Analytics, among others, is based. UTM parameters, which are added to URLs, allow to verify the source of user acquisition. If a person, after accessing Steam from a link containing UTM, makes a conversion (a purchase, addition to a wishlist) within 3 days from clicking, the conversion will be attributed to the last link clicked.
We can find data about the page views and conversions attributed to each link containing UTM parameters in the Steamworks panel of our game, under Marketing and Visibility > UTM Analytics. In the report, we can also find information about the counted page views. This metric tells us how many users logged into Steam during the selected time period in the browser in which they opened the link.
View from the Steamworks panel — UTM Analytics
If a user follows the link on their phone, logs into Steam from their phone and completes the conversion on their computer, the conversion will be credited to our link. If a user does not sign in after clicking on a link on their phone but adds the game to their wishlist on their computer, when they access the game from another source, the conversion will not be credited to our link.
Page view data updates hourly, but key conversions appear in the reports with at least a 1-day delay. During our sales campaigns, the wait time for final results for a particular day extended up to 5 days (which was also affected by the 3-day conversion window mentioned earlier).
What does UTM Analytics change in games marketing?
First of all, we have the possibility to measure the effectiveness of particular marketing channels, the effectiveness of campaigns in target groups and advertising creations in these groups. By appropriately describing links, we can also compare the performance of our advertisements depending on the device on which they are displayed.
So far, to assess the effectiveness of marketing activities, we and our clients have looked for a correlation between expenditure and the number of wishlists or sales in a given market. Seeing the positive impact of a campaign on results, a natural step would be to increase the media budget. However, it’s difficult to optimise budget allocation when we don’t have data on which audience group or ad is performing best.
The data collected by Steam is not perfect. The platform does not use advanced user tracking or third-party cookies, which it emphasises very strongly in its message about its new analytics tools. Nevertheless, we receive information that allows us to optimise our actions and allocate our budget based on real results. This translates directly into an improved ROI of marketing activities, which is crucial for the business of game producers. As an agency specialising in performance marketing, it allows us to do what we feel best at – test, analyse and introduce changes that affect the effectiveness of our clients’ campaigns.
The questions we get answers to with Steam UTM Analytics
Which channel works best?
In the campaign to build the wishlist, we tested the effectiveness of 3 channels — Facebook, Twitter and Reddit.
Depending on the title or creation, the effectiveness of the channels can of course vary. The promotion I refer to here concerned a title with several thousand wishlists, whose communication was targeted at a predefined niche among gamers. Facebook, which offers the most advanced targeting possibilities, performed best in the implementation of this type of campaign. We paid 1.5x more than on Facebook for a wishlist on Reddit. The weakest in our test was Twitter with almost 4.5x the cost of obtaining a wishlist.
Which creations obtain cheaper wishlists?
Until the introduction of Steam UTM Analytics, optimisation of paid game promotion campaigns was largely limited to ad panels. Ad sets and ads with the highest cost per click were turned off by us, and the budget was directed to creations that clicked more cheaply. We now have the opportunity to test the extent to which this approach is effective.
During the campaign encouraging people to wishlist the promoted title, we used video and static creations. As expected, the video ad aroused more interest among users, gaining clicks at an average cost of 30% lower than the static ad. Our assessment of creations and budget allocation changed after analysing data from the Steamworks panel. We checked how many wishlists were assigned to each of the ads and compared this to the expenses. It turned out that despite the higher cost per click, static ads were able to acquire additions to wishlists at a lower cost. By an average of 35%!
How do conversion costs differ between countries?
The choice of countries to which we target ads depends on, among other things, the genre of the game or the media budget. When choosing, it is worth bearing in mind the higher cost of reaching gamers from the United States or Great Britain in media channels. The cost per click, which is several dozen percent higher, cannot always be compensated for by a higher conversion rate. Steam UTM Analytics allows us to see whether we can spend our budget more effectively in smaller markets.
The campaign showed that the cost of acquiring a wishlist between premium markets (North America, Europe) can differ by up to several times.
UTM Analytics data has helped us to align our budget with the return on spend in each country, as well as respond to changes in efficiency over time.
How do conversion costs differ between devices?
Another question we regularly ask ourselves when implementing campaigns to promote desktop games is about devices. Is it worth paying several times more per ad click to reach only users who browse Facebook on a computer?
This narrowing, combined with appropriate targeting, significantly increases our chances of reaching people who are regular gamers. Additionally, desktop users are more likely to be logged into Steam, which shortens the conversion path.
However, the described solution also has disadvantages. Restricting activities only to desktop computers requires much more active campaign management. The number of users is lower, which affects the frequency with which recipients see our ads. Therefore, we have to update creations more frequently in order to maintain their effectiveness.
The cost of acquiring a wishlist from ad sets targeting all devices was on average over two times higher than for those narrowed down to computers. After a few weeks of operation, as expected, we observed a significant drop in the effectiveness of the most frequently displayed ads.
How does Steam UTM Analytics affect the effectiveness of marketing efforts?
During the first 28 days of the campaign, wishlist growth in the five countries we targeted was 161.62% compared to the previous period. At the same time, the growth in the remaining countries was 37.63%.
The key to assessing the impact of the new tools on optimisation is to compare the first and fourth weeks of the campaign. In the fourth week, the cost of acquiring a wishlist assigned to the campaign was 22.16% lower, while the media budget was 317.95% higher. We can thus see how three weeks of intensive testing and optimisation using Steam UTM Analytics data translated into increased campaign effectiveness.
- Increase in the number of wishlists by 161.62%.
- Wishlist acquisition cost decreased by 22.16% while media budget was 317.95% higher
What else is not working properly?
Not all conversions are attributed to the source.
The data presented in the UTM Analytics view does not include all conversions that campaigns are responsible for. This is due to, among other things, the need to log in to the browser where the link opens in order for Steam to attribute a conversion made on another device.
During the course of our activities, we were able to draw conclusions about the difference in counts mainly in the markets which generated 3-4 wishlists per week before the campaign. After the launch of the campaign we observed a jump in wishlists the following week by several dozen. What proportion of all „excess” wishlists did Steam report by linking them to our links? Only between 30 and 50%.
Steam’s counts during ongoing campaigns only took into account 30 to 50% of the wishlists that added value from the campaign.
Errors in the operation of the „page view counts” metric
At the end of April, there were errors in the results concerning counted page views. Data for a few previous days was overwritten, and from that day on, the results of the counted page views were many times divergent from reality. For example, the number of counted page views for a given campaign was often higher than the number of clicks generated by this campaign.
Errors in data presentation charts
The conversion graphs for the time interval selected by the user do not show the correct data, or even the correct interval. Admittedly, this is the least significant problem from a campaign perspective. However, it shows that the tool is not yet perfected. Being able to see a visualisation of the results directly after entering Steamworks would give a quick overview of how the campaign is going.
The screenshot below shows the data visualisation in the UTM Analytics view, after selecting the time interval April-May 2021. The time interval has been changed to January to January. The data for the selected interval also does not match the results reported by Steam.
Screen from UTM Analytics view – Activity over time
Another limitation that makes it difficult to work with the data directly in Steamworks is the way the results summary in the UTM Analytics view works. The results are presented in a separate row for each unique UTM combination. This means that if we promote 5 ads with different links in a set, we will see 5 rows in the summary. We also can’t display them like in the Facebook Ads dashboard, taking into account the overarching level in the campaign structure (ad set, campaign) to compare the performance of two ad sets. To do this, we would have to add up the results for all the ads running in them.
The function of downloading files in csv format comes in handy when analysing large data sets. It can be found on the right-hand side, directly above the breakdown.
Screen from the UTM Analytics – Breakdown view
- Steam UTM Analytics allows you to lower conversion acquisition costs and allocate marketing budgets more effectively when promoting games in paid media channels.
- The tool remains in beta, as evidenced by errors in data presentation and counting of some metrics.
- Due to the mechanics of how it works and the lack of advanced tracking, a significant proportion of conversions are not attributed to the acquisition link.
- We get data that allows us to optimise actions and positively impact their ROI. However, they do not show the full impact of the campaign on results. We need to keep this in mind when calculating the wishlist cost, sales or ROI of paid campaigns.
FormUp is an experienced team of experts specializing in the field of performance marketing. They know their sector from the inside, but what is the most important – they can successfully navigate the uncharted waters.
We work with a great team – an excellent, professional approach, high standards, and above all, enthusiasm – there’s no fear of the unknown, there’s always a new idea or a custom solution ready for implementation immediately. We make a really good team! Thanks for everything!
Have a question or want to find out more?
Contact our expert today and discover how our team can help amplify your digital marketing!