TL;DR: In Philippine alcobev, the shelf, the visi-cooler, and the back-bar at the point of purchase decide the pour. Yet most brands still audit it blind: promodisers snap photos, drop them on Viber, and nobody sees the gap until the campaign is half over. Manual audits cannot keep pace, and ML-based shelf tools stalled on pre-training, an 80% accuracy ceiling, and costs that forced brands back to sampling. BeatRoute's VM Audit AI Agent returns five visibility metrics from one shelf photo, in real time, inside the rep's visit, so a brand can defend facings and chillers while the campaign is still live.
Shelf visibility is the real-time picture of a brand's facing count, visi-cooler placement, POSM deployment, and back-bar shelves across its outlet network. In alcobev, it is the primary sales lever at the point of purchase, especially in a market where liquor bans on election day, Holy Week restrictions, and local ordinances mean a brand cannot always control secondary sales. The brand that owns the chiller and the facing wins the pour. The brand that does not loses it to whichever competitor occupied that space first.
That fight is fiercest where most Filipino drinking money is actually spent: the traditional trade. The Philippines has roughly 1.3 million sari-sari stores (DTI), and alongside them sit the carinderia, the beerhouse, and the videoke bar where a cold bottle is bought one at a time. Whoever's malamig beer is in the front of the chiller, and whose streamer hangs over the door, takes the sale.
Yet across the industry, brands run this lever blind. The VM team designs planograms and POSM at head office. The promodiser or merchandiser visits, takes photos, and fills a checklist. Those photos sit in a shared drive or a Viber chat nobody is watching closely enough. By the time a non-compliance is spotted, if it is spotted at all, the campaign window is half over and the competitor has taken the facings.
The brand never really knows, in time, whether its planogram was executed, what its share of shelf looks like this week, where POSM kits are missing, or which outlets the competition just walked into. Poor-visibility stores stay that way. And the brand keeps losing sales it cannot trace back to the shelf.
This piece breaks down why ML-based shelf audit tools failed before they scaled, and what it actually takes for alcobev brands in the Philippines to turn shelf visibility into real-time intelligence.
Why did ML-based shelf audit tools fail to scale in alcobev?
They failed because manual audits could not keep up with how fast shelves move, and the ML tools that promised a fix broke on training, accuracy, and cost the moment they went live. The operation was scattered across too many tools, too many people, and too many steps to move at the speed the shelf required. Here is where each attempt fell apart.
Pre-training was a non-starter
The model needed thousands of shelf photos per SKU just to get going. In alcobev, that is a moving target. Seasonal packs, festival editions, promo bundles, and market-specific labels change constantly. By the time the model was trained on a SKU set, that SKU set had already changed. Brands were always playing catch-up on a shelf that would not stand still.
Accuracy capped at 80%
Every time the system flagged a non-compliance, someone still had to check it before anyone could act. That is the same manual step brands were trying to remove. An 80% ceiling is not a field tool, it is a second review queue.
The cost made scale impossible
Processing every shelf photo was expensive. So brands did what they always did: audit a sample of outlets, on a periodic cycle. The technology changed, but the economics did not. In a country where coverage means three separate networks across Luzon, Visayas, and Mindanao, a sample is a guess dressed up as data.
Across every pilot, the workflow changed but the core problem did not. Brands were still getting shelf visibility intelligence after the campaign window had already closed.
What does BeatRoute's VM Audit AI Agent do differently?
It does not improve on the traditional approach. It removes the constraints that made it fail. BeatRoute's VM Audit AI Agent sits inside the SFA workflow and is triggered during the rep's standard outlet visit. The sales rep or moving merchandiser only takes a shelf photo. The agent processes it and returns structured visibility data into the visit record in real time. No end-of-day mag-encode. No manual review queue. Here is how it closes each gap.
No training cycles holding up deployment
The VM Audit AI Agent needs no image library and no SKU-by-SKU model training. Deployment takes about five minutes. In alcobev, where packaging and market-specific labels shift across seasons and regions, that is the difference between a tool that works on day one and a pilot that dies before it scales.
Accuracy the field can actually act on
The agent delivers 92%+ accuracy. That is the threshold where a flagged planogram non-compliance stops needing a second pair of eyes before anyone moves. National and area sales managers get alerts they can act on directly, not a queue of unverified flags to work through first. The manager gets truth, and the promodiser gets a visit that closes itself instead of homework to encode after dark.
Economics that make full-network coverage possible
At roughly 40% lower cost per image than traditional ML-based shelf audit tools, the VM Audit AI Agent makes auditing the whole network viable for the first time, not just a sample. Brands stop checking a subset of outlets and hoping it is representative, and start seeing every sari-sari chiller and every beerhouse back-bar on the route. Coverage stops being a sample the brand hopes speaks for the rest.
Five visibility metrics from one shelf photo
One shelf photo returns five visibility metrics the brand previously had no reliable way to track in real time:
- Availability: is the SKU actually on the shelf right now, or did it go out of stock between visits?
- Share of shelf: how much space does the brand hold versus competitors, in the chiller and on the back-bar?
- Facing count: how many units are visible from the front, and does that match the planned execution?
- POSM presence: was the tarpaulin, streamer, or chiller branding actually placed, or is it missing from the outlet?
- Planogram compliance: does the outlet match the VM team's planned shelf layout?
That is five answers, from one photo, returned instantly. Not in a shared drive. In the visit record, where someone can act on it. And because it runs on any Android, even low-end devices, online and offline, a promodiser in a provincial dead zone gets the same audit as one in Makati.
Competitor visibility as structured data
Competitor facing count, new placements, visi-cooler takeovers, and back-bar shifts are all captured during the same visit. Finding out a competitor gained shelf space across your territory this week, while the campaign is still live, is a completely different conversation from finding out at the quarterly debrief when there is nothing left to do about it.
How Philippine alcobev brands turn this into a field advantage
Real-time shelf data only matters if it reaches the person who can fix the outlet before the window closes. Because the audit lives inside the visit, an area manager in Cebu can see a chiller takeover in Davao the same day and reroute a promodiser to defend it, not read about it a month later. BeatRoute is a global platform tailored for the Philippines, with a track record of working well here, which is why major Philippine brands like San Miguel and Monde Nissin run on it. Global scale, local proof, and a field team that is home before dark instead of encoding photos late into the night.
Conclusion
A planogram compliance gap spotted in week one of a four-week campaign can still be fixed. The same gap spotted weeks later is just delayed visibility.
That is the difference BeatRoute's VM Audit AI Agent makes. Non-compliance gets corrected while the campaign is live. Competitor gains get flagged while there is still shelf space to defend. Coverage stops being a sample the brand hopes is representative and becomes a signal across the full network, sari-sari by sari-sari, beerhouse by beerhouse.
Brands still running VM on periodic spot checks are losing facings they cannot see being lost. BeatRoute's VM Audit AI Agent puts the brand back in the visibility battle, outlet by outlet, in real time. Book a PH-tailored demo to see how it works inside your existing field execution workflow.
Frequently asked questions
What is shelf visibility in alcobev?
Shelf visibility is the real-time picture of a brand's facing count, visi-cooler placement, POSM deployment, and back-bar branding across its outlet network. In Philippine alcobev, it is the primary sales lever at the point of purchase, from the sari-sari chiller to the beerhouse back-bar. Visibility decides what a drinker reaches for, so it is the one retail execution metric no brand can afford to leave unmeasured.
How do alcobev brands track shelf visibility across sari-sari stores and beerhouses in the Philippines?
Most still rely on manual photo audits sent over Viber and consolidated in spreadsheets. Photos arrive days late, coverage is sampled, and by the time a gap is confirmed the campaign has moved on. Brands closing this gap use automated shelf audit tools embedded in the rep's visit workflow that return structured data instantly, outlet by outlet.
How does BeatRoute's VM Audit AI Agent improve shelf visibility?
The rep takes one shelf photo during a standard outlet visit. The agent returns five metrics: availability, share of shelf, facing count, POSM presence, and planogram compliance. It delivers 92%+ accuracy with zero pre-training and roughly 40% lower cost per image than traditional ML-based tools, so full-network coverage becomes viable instead of a periodic sample.
Does the VM Audit AI Agent work offline on low-end Android in the provinces?
Yes. It runs on any Android, even low-end devices, online and offline. A promodiser in a provincial signal dead zone can complete the shelf audit during the visit, and the data syncs once a connection returns, so patchy coverage across Luzon, Visayas, and Mindanao never becomes a blind spot.
Can BeatRoute track competitor shelf and chiller presence in alcobev?
Yes. Competitor facing count, new placements, visi-cooler takeovers, and back-bar visibility shifts are captured during the same outlet visit. Teams get competitor shelf intelligence inside the campaign window, not at the quarterly review when there is nothing left to defend.
