What the report measures
Steam's traffic analytics answer three questions:
- Where inside Steam did the game appear?
- How many impressions did each placement produce?
- How many visits to the product page followed?
The report is anonymous. It shows placement performance, not a player-level journey. It also does not prove which external campaign caused a later Steam recommendation.
Use a simple derived measure for each row:
impression-to-visit rate = visits ÷ impressions × 100
If a placement records 20,000 impressions and 400 visits, its derived rate is 2%.
Do not compute it for every row: Valve notes that some platform impressions, such as the pop-up telling a player that a friend is in a game, do not link to the store page at all, so no click-through rate is measurable for those.
Where the rate does apply it is useful for diagnosing a capsule or audience fit, but Valve says conversion rate itself is not an algorithmic visibility input.
Read reach and response separately
| Pattern | Likely meaning | Next check |
|---|---|---|
| Impressions down, rate stable | Steam showed the game less often | Stage eligibility, tags, player response, release competition |
| Impressions stable, rate down | The placement still exists but earns fewer clicks | Capsule, positioning, neighbouring titles |
| Both up | More reach and stronger response | Identify the event and preserve the creative/source context |
| Visits up without Steam impressions moving | External or unclassified traffic may be responsible | UTM and external campaign records |
Do not combine every feature into one blended trend before looking at rows. New Releases, tag pages and personalized recommendations have different eligibility and audiences; a gain in one can hide a loss in another.
A weekly reporting workflow
- Record impressions and visits for each placement. Valve does not document an export control for this report, so plan on copying the figures out by hand.
- Calculate the impression-to-visit rate for each feature.
- Compare with the previous four-week median, not only the previous day.
- Annotate releases, festivals, discounts, capsule changes and Update Visibility Rounds.
- Separate Steam-native placement changes from external traffic.
- Write one diagnosis and one action, such as "search impressions stable, click rate fell after capsule change; restore the control."
Valve generates these reports for each app and also summarises them across the whole set of games on a Steamworks partner account, so multi-title publishers get an account-level view as well as the per-product one.
Reach the account-level view by selecting your partner account name in the blue text on the left, just below "Source". Keep the placement definitions consistent across titles before comparing them.
What the report cannot prove
It cannot identify the person behind an impression, tell you the unpublished ranking weight of a signal, or attribute every visit to one external campaign.
The report also cannot turn a correlation, such as more external activity before more Steam impressions, into proof that traffic directly caused visibility. Valve explicitly says raw store-page traffic is not a visibility factor.
For external-channel attribution, use tagged links and a separate measurement system. Attribution connects ad spend, Steam link tracking and wishlist or purchase outcomes by channel. For Steam-side surfacing, use the Steam discovery algorithm guide.
