Evidence you can follow.
This is an F&B content, visual identity and audience research library. Coimbatore is the first market and SAZZ is the first worked example. It is a sampled library, not a ranking of restaurant quality.
What was inspected
Public profile and post pages, visible captions, and one actual image or video frame for each available post. Full videos and every carousel slide were not assessed. The target is 20 distinct posts per account, including pinned posts and visible collaborations.
Colour and typography
Colour bars are approximate pixel measurements from equally weighted captured images, grouped into 12 named colour families. Photographic scenery, text and captured creator framing all contribute. These are not official brand-palette percentages. Font descriptions are visual classifications; exact font names are unverified.
People and performance
Solid person silhouettes encode relative typical busyness bands, never actual headcounts. Colour indicates suggested persona fit; several colours do not establish audience proportions. Custom time ranges use overlaps with proposed visit windows. Morning matches additionally need a breakfast or morning-routine cue from the account; historical one-off events are excluded, and opening-hour sources and any conflicts are available in the time-range control. Unknown schedules use recorded chart windows, explicitly labelled. The six distinct realistic persona portraits are AI-generated fictional characters, not customers or reviewers. Age groups, visit days and time windows are hypotheses to test; they do not represent observed ages, entry restrictions or verified opening hours. Per-post “Likely appeal” comes from explicit caption cues, with its basis shown. Venue matches can also use reviewed creative cues. Audience cards describe inferred customer needs and suggested opportunities. They are not measured demographics, reach, preferences or purchases. Likes do not tell us why somebody liked a post. No comparative engagement ranking is made.
Busyness & footfall
Typical Google Popular Times charts are captured through public Search. The heatmap uses rendered bar heights, normalized to each place’s own highest captured weekly bar. Dark cells show relative busyness, not headcounts, capacity or comparative visitor totals. Disabled/unshown hours are unknown. Live overlays are excluded. Persona outlines are suggested content-based visit windows; actual persona/age counts remain unknown.
Google Maps
The full SAZZ audit contains five accessible Google Maps reviews and six customer photos. The first Indiranagar batch adds three platform-selected reviews and one customer photo per café. Several review excerpts are partial; the sample is not representative. Each place has attribution, source links and collection dates. Brand-posted testimonial artwork is selected marketing content, not an independent review sample.
Places and account scope
Map pins use branch listing coordinates from linked District or Google Maps sources. Choose an area to keep local comparisons together. Shared Instagram feeds are labelled and can include other branches. Annalakshmi’s Instagram address differs from its mapped listing and needs branch confirmation. Bazaar uses its hotel’s feed; Little Soi uses a multi-city account. The French Door’s official linked account was unavailable.
Playback and offline use
Instagram embeds require internet and may ask for login or be unavailable if a creator restricts embedding. Original post links are always provided. Saved previews, notes and the location plot remain available in the offline package; street-map tiles need internet.
Attribution
Post images belong to their original creators and are displayed as attributed research references. Map: OpenFreeMap · OpenMapTiles · OpenStreetMap contributors. Map software: MapLibre GL JS; licence included.
Open the complete SAZZ audit →