Mobile ad tech firm Celtra has released its Mobile Display Ad Performance Report for Q2 2015, based on ads which run on the Celtra platform.

The report reveals that banner format impressions are gradually increasing over time, and already represent almost a third (30 per cent) of all quarterly requested impressions. The share of impressions in the web browser environment hit an all-time high at 44 per cent.

The report includes viewability benchmarks for the first time. 63 per cent of impressions were viewable by the IAB viewable impressions definition, in which at least 30 per cent (for large size ads) or at least 50 per cent of the creative (for all other ads) was displayed to the user for at least one continuous second. The highest IAB viewable rates were for Interstitial format (70 per cent IAB viewable rate).

The report also looked at ad engagement across different ad formats. Interstitials, due to their full screen size, reported the highest engagement rate at 3.9 per cent. The ad engagement (expansion) rate for the expandable banner format (engagement with the banner unit), came in at 0.76 per cent, the highest to date. Engagement time (previously referred to as Time on ad unit) remained steady at around five seconds for interstitials, and was five seconds lower than in Q4 on the expanded unit. The expanded unit engagement rate for the expandable banner format increased by 27 per cent compared to the previous quarter.

The report also reveals big differences in video completion rate depending on the video play setting (user-initiated or auto-play). For banners with auto-play video, the completion rate is higher than for those with user-initiated video (54 per cent versus 31 per cent). But for the expanded unit within an expandable banner, the opposite is true. Here, the video completion rate for user-initiated videos is 59 per cent, compared to 26 per cent for auto-play videos.

The data sample for Celtra’s Quarterly Benchmark Report is composed of ads that start, end and run on the Celtra platform during a specific time period. Ads that are included were part of campaigns with a minimum of 100,000 impressions, each ad running on a minimum of 50,000 impressions in the defined time period. All ads without tracking of events are excluded from the sample.

After the initial database is set, Celtra firstly preforms random sampling of campaigns inside the set of defined verticals. This is followed by a random sampling of ads by ad format and random sampling by ad features, in order to acquire a sample with a balanced ratio of ads from chosen verticals, different ad formats and feature groups. Lastly, Celtra filters the sample to ensure there is no more than one creative per ad format from the same campaign included, in order to avoid duplicating very similar values and thus compromising the final analysis results.