How to Measure Proximity Campaigns
Measuring a proximity campaign means tracking a chain of events that starts with a physical signal and ends, ideally, with a recognisable action. The chain has distinct stages, and each stage introduces a gap where data can be lost or misinterpreted.
The first stage is detection: a device receives a Bluetooth broadcast, reads an NFC tag, or scans a QR code. Detection alone tells you very little. A phone passing a beacon in someone's pocket registers a detection event, but the person may be unaware of it entirely. Counting detections as "impressions" is a common mistake that inflates figures and erodes trust in the data.
The second stage is notification delivery. For beacon-triggered campaigns, this depends on the user having Bluetooth enabled, location services active, and the relevant app installed with notifications permitted. For QR and NFC, delivery is user-initiated, so the gap between detection and delivery collapses to a single tap. This is one of the practical reasons QR and NFC often produce cleaner measurement data than beacon-only campaigns.
The third stage is engagement: the user opens the notification, views the content, or interacts with the landing page. The fourth is conversion: the user takes a defined action such as making a purchase, signing up, or downloading a resource. Not every campaign needs to reach the fourth stage, but you should define which stage constitutes success before the campaign launches.
Second, you need a representative baseline period with the infrastructure active but the campaign content suppressed, so normal footfall can be compared with the campaign period. The necessary duration depends on trading patterns, venue schedules and sample size; one week is not automatically sufficient for every site.
Defining Campaign Success Metrics
The right metrics depend on what the campaign is actually trying to achieve. A museum wanting to increase dwell time at a specific gallery needs different measures from a retailer pushing a time-limited offer. Forcing every campaign into the same metric framework produces numbers that look tidy but mean little.
For awareness and footfall campaigns, useful metrics include the number of unique devices detected in a zone, the change in dwell time compared to the baseline period, and the ratio of return visits. For engagement campaigns, notification open rate, content view duration, and tap-through rate to secondary content are more relevant. For conversion campaigns, the metrics that matter are the ones tied directly to a business outcome: redemptions, transactions, sign-ups, or whatever the defined action is.
A practical way to structure metrics is to separate leading indicators from lagging indicators. Leading indicators, such as notification delivery rate and open rate, tell you quickly whether the technical delivery and message relevance are working. Lagging indicators, such as conversion rate and revenue impact, take longer to accumulate but reflect actual business value. If leading indicators are strong but lagging indicators are weak, the problem is usually in the offer or the content, not the infrastructure.
One frequent error is treating notification delivery rate as a success metric in its own right. A high delivery rate simply means the beacons are broadcasting and the app is receiving signals. It does not mean the content is relevant, the timing is appropriate, or the user is interested. Another error is comparing proximity metrics directly with digital advertising metrics without adjusting for the fundamentally different context. A two percent tap-through rate on a push notification triggered by a physical zone may represent excellent performance, whereas the same figure in an email campaign would be poor.
As an illustrative example only: if a baseline week shows an average of 400 unique devices per day in a zone, and a campaign week shows 520 unique devices with a notification open rate of 14 percent and a conversion rate of 3 percent on opened notifications, you can calculate the incremental conversions attributable to the campaign. But this figure remains an estimate unless you can isolate the campaign as the sole variable, which is rarely possible in a live environment.
Attribution Challenges in Proximity Marketing
Attribution in proximity marketing is fundamentally harder than in purely digital channels because the physical world introduces variables you cannot control. The core problem is establishing a causal link between a person being in a specific location and subsequently taking an action, when multiple other factors may have influenced that action.
Consider a customer who receives a discount notification near the footwear aisle and later buys shoes at the till. The proximity system records the notification delivery and the app records the purchase, but you cannot be certain the notification caused the purchase. The customer may have already decided to buy shoes before entering the store. They may have seen a poster, spoken to a staff member, or been influenced by an entirely separate online campaign. Attributing the sale entirely to the beacon notification overstates the campaign's effect.
Multi-touch attribution models attempt to distribute credit across several touchpoints, but they require a level of data integration that most proximity deployments do not have. In practice, most proximity campaigns rely on last-touch attribution within the app environment, which means any interaction that happened outside the app, or before the app was installed, is invisible.
Unidentified or aggregated detection creates a measurement gap. If the system cannot legitimately connect an interaction to a person or transaction, it can support zone-level counts or trends but not individual attribution. Do not call records anonymous merely because names are absent: persistent identifiers and linkable event histories may still be personal data. Report what can be linked, what cannot and what safeguards were applied.
Practical ways to improve attribution confidence include using control zones where no campaign content is delivered, running time-limited campaigns so you can compare narrow windows, and pairing proximity triggers with unique codes or QR-based redemption paths that create a direct, traceable link between the physical interaction and the conversion. None of these methods eliminates the attribution problem, but they narrow the uncertainty.
Analytics Dashboards for Proximity Campaigns
A useful analytics dashboard for proximity campaigns needs to bridge two very different types of data: signal-level data from the physical infrastructure and business-level data from the campaign outcomes. Dashboards that show only one side of this equation are common and largely unhelpful.
Signal-level data includes beacon detection counts, RSSI values, battery status, and zone entry and exit timestamps. This data is necessary for operations teams to verify that the infrastructure is working, but it means very little to a marketing manager trying to evaluate campaign performance. Business-level data includes notification delivery and open rates, content interactions, conversion counts, and revenue figures. This is what marketing teams need, but without the signal data underneath, they cannot tell whether a poor result is caused by a campaign problem or a hardware problem.
The most practical dashboards present a summary view oriented towards business outcomes, with the ability to drill down into signal health when something looks wrong. If conversion drops suddenly in a specific zone, the first question should be whether the beacons in that zone are still broadcasting at the expected power and interval, not whether the creative needs to be redesigned.
When evaluating a dashboard from a provider or integrator, check whether it can filter by zone, by time window, and by device platform. Check whether it shows unique devices rather than raw event counts, because a single device lingering in a zone can generate thousands of detection events that distort averages. Check whether baseline comparisons are built in or whether you will need to export data and calculate them yourself. Check how far back the data is retained and whether you can export it in a standard format for your own analysis.
Common dashboard failures include displaying real-time detection maps that look impressive but provide no actionable insight, showing aggregate averages that mask zone-level variation, and presenting notification delivery rates without any context for what a normal rate looks like in that specific environment. A dashboard that cannot answer the question "is this zone performing differently from last week?" is not doing its job.
Reporting to Stakeholders
Different stakeholders need different information from proximity campaign reports, and compressing everything into a single summary slide usually satisfies no one. Operations managers need to know about infrastructure health, battery status, and any zones where detection rates have dropped. Marketing managers need campaign-level performance against objectives. Finance and senior leadership need a clear connection between the proximity investment and business outcomes, expressed in terms they already use.
The most effective reports lead with the campaign objective and whether it was met, then provide the supporting evidence, then acknowledge the limitations. Stating "the campaign drove a 12 percent increase in dwell time in Zone C compared to the baseline week, though we cannot fully isolate the proximity trigger from the concurrent in-store signage change" is more credible than presenting the 12 percent figure without qualification. Stakeholders who discover unmentioned caveats later lose confidence in the entire measurement system.
Report cadence should match the campaign cycle. A week-long pilot warrants a detailed report within days of completion, while an always-on proximity installation in a museum might be reported monthly with weekly operational checks. Avoid the trap of producing reports simply because the calendar says to. If nothing meaningful has changed since the last report, say so rather than repackaging the same figures with different formatting.
When presenting proximity data to stakeholders who are not familiar with the technology, avoid jargon and translate signal terminology into plain language. "Average RSSI in Zone D was minus 62 dBm" means nothing to most people. "Devices in Zone D were detected reliably at an average distance of approximately three to four metres, which is within the expected range for this placement" conveys the same information in a usable form.
A practical reporting structure for a single campaign includes: the objective stated in one sentence, the baseline period and its key figures, the campaign period and its key figures, the calculated difference, the known variables that may have influenced the result, the proportion of detected devices that were anonymous versus identified, and a clear statement of what the data can and cannot support. This structure forces honesty and gives stakeholders what they need to make decisions about future campaigns.
