Validate taxi DOOH exposure, sync audiences with privacy safeguards, set bid and sequence rules, run a small pilot, and measure lift.
Taxi ad playback does not prove someone saw your ad. I’d start programmatic DOOH retargeting with three checks: valid screen records, permitted audience matching, and confirmed support for follow-up ads.
The goal is to connect a possible taxi-screen exposure to later mobile, web, app, or connected TV ads - not target someone just because they were nearby.
Here’s what I’d check before spending:
My approach: start with a small pilot, fix gaps, and <u>change one variable at a time</u> before expanding.
Taxi DOOH Retargeting: From Playback to Measured Lift
Once you’ve defined the exposure signal, connect the screen, buying system, and audience sync process. Use two linked paths: screen → CMS → SSP/exchange → DSP for activation, and playback records → approved matching service → DSP for retargeting.
The CMS manages devices, schedules, content, and proof-of-play. The SSP packages inventory, while the DSP checks bids against campaign rules. Document each connection’s API, owner, IDs, event names, and permissions so every platform passes clean IDs and timestamps downstream.
Enroute View Media’s DigiStreamView platform can manage vehicle screens, scheduling, proof-of-play, reporting, analytics, location-and-time targeting, and programmatic integrations. These tools provide the screen-side base for retargeting.
Before booking inventory, get written confirmation of supported SSPs and DSPs, API access, reporting exports, campaign ID passback, and permitted data use. Programmatic integration alone does not confirm an identity-matching or retargeting link. Next, make sure every platform uses the same data definitions.
Use a shared schema for screen ID, vehicle or fleet ID, market, approved geography, operating hours, screen type, resolution, slot duration, and estimated impressions. Pass campaign, creative, bid, and deal IDs through every stage: bid request, win notice, creative handoff, playback, and reporting passback. Keep inventory, bids, and ad delivery separate, and document how impression estimates are calculated.
Before launch, confirm platform access, define event names and required fields, assign IDs, and set time-zone and daylight-saving rules. Set reporting cadence, latency, retention, and deletion periods, too. Test bid requests, creative responses, duplicate events, late logs, offline playback, and clock drift. With these records standardized, audience activation can use them without passing raw location data downstream.
A playback record describes a screen event - not an identified viewer. Send validated events to an approved service for permitted time-and-location matching. Before creating segments, require the matching method to account for vehicle movement, route variation, dwell time, and location uncertainty.
Sync only permitted pseudonymous identifiers to the DSP; pseudonymous does not automatically mean anonymous. Document data rights, sensitive-location exclusions, minimum audience sizes, and expiration rules. Use only the location precision needed, limit retention, and pass suppression and deletion instructions downstream. Get explicit contractual permission before household or cross-device activation.
With screen IDs and audience sync in place, set bid rules for each inventory type and follow-up channel. Follow one workflow: trigger → bid rule → sequence → asset match.
Set eligibility before adjusting bids. Keep rooftop LED and in-taxi tablet rules separate: visibility at passing distance and passenger interaction serve different goals. Give each inventory type its own bid ceiling and budget. Set a CPM floor and campaign-level bid ceiling to keep optimization within the approved rate.
Use the same exposure segment to control bid eligibility and follow-up message order. Apply modifiers first, then enforce the final bid ceiling and fallback defaults. For every campaign, define trigger, recency, cap, and expiry rules before launch.
| Rule area | Required filter | Bid modifier | Hard exclusion |
|---|---|---|---|
| Geography and routes | Approved ZIP codes, geofences, or routes | Prioritize high-value neighborhoods | Outside approved boundaries |
| Daypart and availability | Supported local hours; active, compatible screens | Favor stronger estimated dwell time | Offline screen or unavailable slot |
| Audience signal | Minimum modeled audience density | Increase bids where modeled density exceeds the campaign threshold | Missing required measurement coverage |
| Weather and events | Supported, timely signals when required | Adjust for approved weather, venue, or event conditions | Missing required signal without an approved fallback |
| Price and budget | CPM floor and campaign-level bid ceiling; separate daily and total budgets | Adjust only below the final bid ceiling | Floor above ceiling or exhausted budget |
| Follow-up audience - separate rules | Permitted segment, sufficient exposure confidence, active recency window | Prioritize eligible recent exposure | Opt-out, expired membership, or frequency cap |
Once bid limits are set, give each channel a role: taxi screens for the first impression, mobile for the reminder, and video or CTV for reinforcement.
Keep campaign IDs, offer codes, destination URLs, market definitions, and message themes aligned, but change the message at each stage. Include eligible exposure segments. Where appropriate, suppress users who have opted out or reached the frequency cap. Treat the sequence as a planned order, not a person-level guarantee.
The ad asset handoff should preserve the campaign ID, offer, and recency logic used in bidding. Check dimensions, aspect ratio, duration, file weight, codec, audio behavior, safe margins, approvals, expiration dates, and applicable destination URLs against each receiving partner’s specifications.
Test rendering and interactions before release. Supply pre-approved fallback assets with the same offer and campaign ID.
| Placement | Role | Format requirements | Timing check |
|---|---|---|---|
| Rooftop LED | Quick introduction | Large type, high contrast, no audio dependence; readable at passing distance | Confirm booked spot and loop duration |
| In-taxi tablet | Passenger action | Touch targets; tested QR codes or links; controlled audio | Match ad content to available passenger dwell time |
| Mobile | Reminder or offer | Supported dimensions, file weight, working destination URL | Enforce recency and frequency limits |
| Video | Reinforcement or explanation | Placement-specific dimensions, codec, captions, audio fallback | Confirm allowed duration and playback behavior |
| Connected TV | Longer explanation | Compatible video, captions, audio mix, player requirements | Confirm accepted spot length |
After assets pass spec checks, pace delivery and measure lift.
After the ad assets are handed off, focus on pacing and measurement. Set separate budgets for taxi exposure, follow-up media, and conversion activity, then adjust them based on early delivery results. Track media, data, platform, ad production, and measurement costs separately as planned, committed, or spent. Release follow-up funds only after synced records pass consent and quality checks.
Monitor pacing daily. Audience expiration, channel frequency caps, delayed device-ID passback, and incomplete cross-device coverage all limit the audience you can spend against. Keep exposure segment definitions consistent from bidding through follow-up spending.
Check pacing at least once each business day, using each market’s local time zone. Compare planned and actual spending, plays, eligible audience size, match rate, follow-up impressions, and frequency. If delivery falls below target at two consecutive checks, pause the line item or require approval before moving more budget into it. When audiences are small, redirect funds to approved taxi reach or conversion activity instead of forcing more impressions.
Once spending begins, measure delivery and lift by exposure cohort. Show delivery, audience sync, and outcomes in one view: taxi exposure first, matched follow-up delivery next, and outcomes last.
Include screen ID, timestamp, location or route, creative ID, duration, and play status in the delivery report. For estimated reach and frequency, disclose audience sources, modeling methods, how repeat views are counted, and missing data. Track interactions, clicks, landing-page visits, and agreed conversions separately. None of these proves attention. Deduplicate conversions across channels and define the attribution window.
Calculate match rate by dividing matched identifiers by consented, eligible exposure records - not screen plays. Measure sync delay from exposure to audience availability. Measure exposure-to-follow-up time from exposure to the first follow-up impression. Report median and upper-percentile delays, along with frequency per matched person or household within a stated window. Matches remain inferred; they do not confirm identity.
| Group | Delivery and outcome measures | What the comparison cannot establish alone |
|---|---|---|
| Taxi-only | Verified plays, estimated reach and frequency, visits, conversions | Taxi’s added impact without a comparable control |
| Taxi-plus-retargeting | The same measures, plus match rate, sync delay, follow-up frequency, clicks, landing-page visits, conversions | Whether better outcomes come from follow-up ads or an audience that is easier to match |
| Control or holdout | The same conversion and outcome measures; document excluded and permitted media | Requires enough participants and protection against contamination |
Use randomized follow-up holdouts among eligible audiences. If that isn’t possible, use matched geographic markets or synthetic-control methods. Document dates, sample sizes, market differences, and conversion windows.
Moving taxis and other media can contaminate groups. Comparing matched users with unmatched users can also introduce selection bias. Last-touch attribution alone isn’t enough: it credits the final tracked interaction, not necessarily the taxi exposure that helped create demand.
| Problem | Check | Corrective action |
|---|---|---|
| Missing play logs | Connectivity, heartbeat, time sync, ID mapping, and ingestion status | Repair the player or feed; backfill only documented records and exclude unverifiable plays from delivery totals |
| Low match rate | Consent, identifier compatibility, eligibility, geography, and denominator | Correct mapping or expand approved taxi inventory or exposure windows without bypassing consent |
| Sync delays | Cross-platform timestamps and batch schedules | Shorten transfer intervals; activate follow-up only after records are valid |
| Rejected assets | Format specifications and policy status | Supply compliant fallback assets and retest before resuming follow-up delivery |
| Uneven pacing | Taxi supply, bid floors, caps, and inventory filters | Adjust taxi eligibility or budget allocation to support follow-up reach before adding spend |
| Excessive frequency | Identity resolution, overlapping line items, and caps | Apply a shared follow-up cap; shorten eligibility or suppress converted users as needed |
Scale only when the entire exposure-to-retargeting chain works - not just the screen. That means validated delivery records, privacy-compliant audience building, and bid limits that fit the available follow-up audience. Start with a pilot to check that these links work together.
The pilot checklist should cover end-to-end SSP/DSP tests, approved taxi and follow-up assets, audience-sync timing and match-rate checks, opt-out and suppression tests, bid and budget caps, baseline measurement, conversion tags, proof-of-play matching, reporting access, and rejection tests.
If the pilot passes, secure partner support and reporting access before increasing spend. Get written confirmation from Enroute View Media on supported U.S. partners, audience-sync support, reporting fields, exports, and account limits before scaling. Then expand fleet coverage, geography, or bid limits one variable at a time.
There’s no fixed budget. Costs depend on your goals, campaign size, target locations, number of vehicles, campaign length, and programmatic CPM strategy. Successful pilots have used 100 to 200 vehicles over 4 to 8 weeks to measure increases in foot traffic and sales.
We recommend a discovery call to discuss your business model and technical needs. From there, we can prepare a proposal that fits your budget.
Retarget at the next available bid opportunity once the vehicle’s GPS indicates entry into your target geofence or your targeting conditions are met. Use daily caps and frequency controls to pace follow-up ads.
The platform logs “proof of play” with an exact timestamp and location. That means follow-up ads can be based on confirmed playback, not just screen uptime. Ad content typically syncs to devices within minutes.
There’s no fixed data volume requirement. It depends on the campaign. What matters is having reliable signals that connect ad exposure to consumer actions, such as GPS-verified proof-of-play and mobile ID matching.
Use incrementality testing with control groups to compare conversion rates for exposed and unexposed audiences. For accurate attribution, prioritize verified exposure logs over estimated impressions.
Need pricing details?
EnrouteViewMedia is a Canadian technology company based in Edmonton, Alberta, Canada. The company was founded by a group of media entrepreneurs and engineering executives. Our mission is developing leading-edge digital media display technologies for taxicab rooftops and taxicab interiors advertising market. Through our easy to use ad delivery and management system, we assist our clients in reaching their customers with targeted exposure. EnRouteViewMedia continuously strives to be the global leader in digital geo-time targeted advertising media systems for modern taxi fleets.