Brand lift measurement for programmatic OOH

• APEX shares the advertiser name, campaign objectives, flight dates, markets, media plan (formats, screen networks or venues, estimated impressions or plays), target audience description, and the KPIs of interest.
• LoopMe runs a feasibility check against market-level minimums. Feasibility is a function of expected exposed device volumes and survey density in each market, so the same campaign behaves differently in different countries. LoopMe returns a study tier recommendation and a campaign-specific feasibility estimate.
• Studies run per market. Where a campaign spans several markets, each market gets its own study read so a shared control group never distorts per-market results.
• A minimum active flight of three weeks is recommended so that exposure and response sample can accrue at a useful velocity.

This is where prog OOH differs from pixel-measured digital channels. A screen has no user-level pixel, so exposure is established by connecting playout to the devices present.

• Playout data. The DOOH SSP or DSP provides play logs covering which screens ran the creative, where each screen is located, and when each play occurred. Any reporting cuts APEX will want downstream (venue type, screen size, creative, daypart, and so on) need to be laid out in these files, mapped screen by screen and play by play, or agreed at a higher level ahead of time. Reporting can only break out what the play logs distinguish.
• Device presence. LoopMe works with location data partners to receive exposure files identifying the mobile devices (MAIDs, and in some cases IP addresses) observed at or near each screen during the play windows. Where the media owner or platform holds its own exposure or app data, direct exposure file delivery is also supported.
• Matching. Exposure records are matched to the LoopMe identity graph on device ID (or, in some cases, IP address, though device ID is best practice).
• The result is a persistent exposed pool of real devices with a high-confidence opportunity to have seen the campaign, refreshed throughout the flight as new play logs and exposure files arrive.

• LoopMe designs a recommended survey from three inputs gathered at intake: the campaign's business objective, the target audience, and any additional learning goals.
• The recommended shape is five questions: an audience profiling question, three brand lift KPIs drawn from the standard funnel (brand awareness, ad recall, message association, brand attribute, consideration, favorability, purchase intent), and a custom insight question tied to the campaign's learning goals. This structure is optimized to deliver the best completion rates on a non-incentivized audience and the richest insights.
• The recommendation is not set in stone. APEX or the end client can adopt it as is, adjust it, or provide fully custom questions.
• Studies can run at different sample levels depending on goals, from light studies delivering a confident topline read through standard and advanced studies with larger guaranteed samples that support real-time dashboards, demographic breakouts, and tactic-level cuts. The right tiering for APEX Connect is something LoopMe and Publicis can define together.
• Guaranteed responses are counted on question one. Because some respondents drop off as the survey progresses, question one carries the largest sample and each subsequent question reports against the sample it retains.
• Survey turnaround is same day to a few days after the brief, with the survey approved by APEX before fielding.

• RISA scans the identity graph to build a control group that is statistically identical to the exposed pool across more than 50 variables, including demographics, geography, device features, behavioral attributes, household income, and content consumption, except that control users have not been exposed to the campaign.
• The control group is concurrent and persistent. It is built and surveyed in parallel with the exposed group throughout the flight, not modeled after the fact, so environmental effects such as news cycles, seasonality, and competitor activity wash out of the lift calculation.

• Exposed and control users are surveyed on LoopMe's own mobile in-app inventory. When a user in either group is about to be served an ad, LoopMe wins the auction and delivers a full-screen survey instead.
• Participation is non-incentivized. Respondents are never paid or rewarded, which removes the professional-respondent and incentive biases inherent in panels and produces answers from people intercepted in natural media moments.
• Exposed users are surveyed within seven days of their most recent exposure to minimize recall decay. The window is configurable by campaign.
• Each user can receive and answer the survey only once for the life of the campaign.
• Surveying runs daily over the course of the campaign. On a short flight of around three weeks, the bulk of surveying typically lands in the final week as the exposed pool reaches full size. The goal is to finalize surveying on the last day of the flight, though the methodology allows fielding to continue through the full seven-day lookback window after the flight ends if needed.
• Surveying continues until the guaranteed response volume is reached, with fielding compliant with GDPR and other applicable privacy frameworks.

• Lift is calculated per question as the difference in positive response rate between exposed and control, reported as relative lift, absolute lift, and lifted users. Lift is reported once at least 10 control and 10 exposed responses exist on a question, and statistical significance once at least 30 and 30 exist, with confidence levels reported at the KPI level.
• Standard reporting is self-serve access to the LoopMe client dashboard, refreshed throughout the flight, with CSV and PDF downloads of all data tables and charts. Reporting access is the same for every study; sample size determines how granular the cuts can be, not what reporting is available.
• For prog OOH, reporting cuts are built from the screens. LoopMe can work with APEX to report out any dimension (venue type, screen size, creative, daypart, region, and so on) provided that dimension is mapped to specific screens, and play times where relevant, in the files shared during setup. With that mapping in place, lift can be broken out by the dimension.
• Outputs also include exposed audience insights (demographic, geographic, and behavioral indexes powered by the DMP) and, where sample allows, an optimization audience built from positive respondents that can be pushed back into activation.
• An analyst-produced post-campaign analysis report is available as an optional deliverable to be scoped in the commercial agreement.

• APEX side: campaign details and media plan at intake, connection to the DOOH SSP or DSP for play logs, survey approval, and sharing results with the end client as needed.
• LoopMe side, fully managed: survey design and programming, exposure file ingestion and matching, control group construction, survey fielding and pacing, analysis, dashboard provisioning, and post-campaign analysis where scoped. A dedicated account manager runs the study and communicates proactively throughout.
• Practical requirement for OLV and display when they join a study: the LoopMe pixel needs to be trafficked on every measured placement, with a data services agreement in place covering pixel-based collection. LoopMe provides pixels same day and supports macro structures for tactic-level breakouts.