These are relevant occasion alternatives, not an Instagram performance league table. Official-site inspection supports the comparisons below; no competitor's feed, engagement or sales was audited. Price equivalence could not be established from current official menus.

Positioning interpretation: SAZZ offers global food and drinks for people arranging an unhurried shared meal or an occasion, positioning itself through kitchen craft and a distinctive dining setting, supported by visible preparation, venue footage and brand-shared press features. Inferred · medium-high from Times of India feature, The Hindu food feature, Welcome to SAZZ, Food and a place to stay and Made for you.

The apparent customer desire is a place worth making plans around, with enough food variety for a shared table. The venue and staff preparation make the experience tangible; food close-ups prevent the account becoming only an ambience catalogue. The bio's promise is much broader than the specific visual evidence. Naming Coimbatore and the food/setting proposition in the bio would help a first-time visitor understand it faster. Inferred · medium; proposed improvement. Instagram profile · Preparing a warm welcome

Price position is unresolved. District lists an approximate ₹1,200 for two, while the older brand-reposted Hindu clipping describes approximately ₹2,500 for two. The Onam campaign's ₹1,200 per person applied to a dated event. These are different contexts and must not become a current menu promise. An upscale experience is a reasonable interpretation of the presentation; a verified price tier or affordability claim is not. District dining reviews · The Hindu food feature · Onam feast reveal

Trust material includes visible staff and preparation, a legible brand mark, press-feature posts and a functioning booking destination. Press posts are self-published reproductions: publication provenance and whether coverage was editorial or promotional were not independently verified. Repeat visits could be encouraged by named dishes, a rotation of drink choices and clearly announced current menus. Those are proposals, not proof of repeat-purchase behaviour.

Paraphrased from the five expanded review cards inspected on 24 October 2026. Ages are as displayed that day. Spend bands are historical self-reports, not a price list. A blank or unreported field is not a negative answer.

Google displayed topic chips for “polite staff” (114), food variety (25), sushi (25), dimsum (16), mango sticky rice (6), big screen (6) and kids’ play area (4). These are Google’s displayed mention counts. We did not read those underlying review sets or establish their sentiment.

Use these as leads for further review research, not as approval rates or verified facility claims. They are kept separate from the five-review sample.

These mutually exclusive categories are this audit’s interpretation. A visit report can contain both praise and a concern; it is counted once. The breakdown covers all 20 posts, not just the examples above.

Currently intended: shared-evening planners, menu explorers and occasion organisers, with an additional community-event collaboration strand. Visibly engaged: the expanded comment sample includes location questions, event-registration keywords, praise, tags and self-reported visits. Purchasing audience: Google Maps and District contain customer self-reports, including groups of 12 and 15; no booking records were supplied. Suggested ideal customers: the current evening/group diner remains the strongest strategic fit; vegetarian-choice planners, celebration coordinators and weekend-lunch planners are useful expansion tests, not proven profitable segments. Creator’s visit · Sharing Reel · Google Maps

These are occasion-based needs, not invented biographies; one person may fit several. Validate with anonymised enquiry themes, reservation reasons, menu availability and voluntary customer feedback. Supplied Insights could test distribution, but audience reach alone still would not verify purchasers.

Customer preferences behind the personas

The People tab now includes six AI-generated editorial illustrations. They show fictional people and possible occasions, not portraits of reviewers, verified SAZZ setups or inferred follower demographics. Prompt and asset manifest. The original PNGs and web JPEGs are included in the package.

In the five full Maps reviews, all five praise food, service and atmosphere. Two describe groups of 12 and 15. Three select quiet/easy to talk; two select moderate noise. All five report ample parking. These are overlapping themes and historical self-reports from a favourable platform-selected sample, not population satisfaction rates or current facility promises.

Google Maps review source. The People tab expands each reviewer’s praise, concerns, rating fields, reported spend, noise, vegetarian recommendation, wait time, parking and booking notes. These fields are also saved in people-evidence.json. Historical spend bands vary widely and do not define a current menu price or customer income.

Post likes and comment likes are different. The creator visit displays 11.9K post likes. One comment praises maki rolls and expresses return intent, with 6 comment likes; it also ends with an unresolved comfort remark. Another explicitly praises food and atmosphere, with 3 comment likes. Three location comments display 1, 3 and 1 likes. A companion-seeking comment displays 4 likes. A separate ambiguous smell remark displays 21 likes. No reply context was inspected; these comments do not establish verified comfort problems, and their likes do not prove agreement. Creator-visit thread

The sharing Reel’s positive comment does not name what was liked. The Pilates invitation contains two keyword responses, one emoji response and an organiser comment; registration and attendance remain unknown. The five-dish Reel’s readable question asks about the content-production team, not an order. The football recap’s negative remark concerns a team, not restaurant service. The People tab shows these examples beside their post images and separates post-level counters from comment-level counters.

Will the current content reach the ideal customer?

Creative fit is promising; actual distribution is unknown. The food, setting and shared-table stories are likely to appeal to evening/group planners. The creator collaboration explicitly names friends, family dinners and dates. Three visible location questions suggest that appealing content still needs clearer directions. The specific five-dish caption is a useful model for answering menu questions.

Do not attach a fabricated reach percentage to these judgments. Validate distribution with location and non-follower reach in Insights, then track menu/directions actions and qualified enquiries by occasion. Bookings require business records.

The recent dominant system is photographic and process-led. The press system is editorial; the Onam system is an intentional seasonal variation. Observed colour descriptions are visual approximations. No exact brand hex values or font files were obtained.

Evidence: Times of India feature, A bowl with a little drama, Preparing a warm welcome, Welcome to SAZZ, Behind every bite, Food and a place to stay, Made for you, Six drinks to discover, Onam memories, Onam feast reveal. All visible-pattern findings are observed, high confidence for these examples; recurrence beyond the sample is unknown.

Measured colour mix

One sampled image per post; equal weight. Figures are rounded and total approximately 100%. The hex values are measured cluster representatives, not approved brand specifications.

Method: One captured cover/frame from each of 20 posts, including collaborations; each resized to 96×96 with nearest-neighbour sampling, equal weight per post. Median-cut quantisation into seven colour clusters. Browser chrome excluded; visible letterboxing and on-image type retained. Values describe sampled pixels, not official brand colours, all video frames or the full feed. Exact font files remain unknown. The visible categories are light sans-serif photo overlays, editorial serif/italic press titles, and heavier seasonal display lettering. The collaboration poster is a separate partner style.

Scores describe inspected execution only: 1 major recurring weakness; 2 frequent weakness; 3 competent but uneven; 4 strong with minor gaps; 5 consistently well executed. They are qualitative reviewer judgments, not performance measurements. No overall score because motion, full content history and product fidelity were not verified.

Reference families: food/process — A bowl with a little drama/Behind every bite/Made for you; drinks — Welcome to SAZZ/Food and a place to stay/Six drinks to discover; press/trust — Times of India feature/The Hindu food feature; event/information hierarchy — Onam feast invitation/Onam feast reveal; visit experience — Preparing a warm welcome/Welcome to SAZZ/Food and a place to stay. Before approving new artwork, place it beside the relevant two or three references, check logo treatment, count/scale of image panels, text zones, food colour and phone-size legibility. Preserve recognisability without copying a layout that does not suit the new information.

Assets required: approved brand files; current menu and dietary wording; latest availability/prices; confirmed reservation line/link and hours; existing original footage; shoot access and staff consent; current celebration/group policies; rights-cleared music; an enquiry/booking measurement owner. Budget and production capacity are unknown.

Snapshot: 24 October 2026. K/M labels are rounded. Reposts are separate from private shares. “0 shown” records the interface, not proof that no comment ever existed. Missing values are unknown.

The creator’s sundowner collaboration has the highest displayed likes (11.9K), comment count (30) and available Reel views (1M) in this 20-post sample. One static post hides its like count, and one Reel’s views were not captured. These are therefore maxima among available counters, not all-time or complete-account records. Collaboration, post age and possible paid support make direct creative-performance comparisons unreliable.

The Times of India feature has 7.4K displayed likes; the football recap has 573K displayed views; the two September/October craft Reels show about 546K–547K. None establishes organic reach, unique people or bookings. Source: each linked post and the Reels tab. Do not combine reposts with private shares or calculate a conversion rate from these public counters.

All 20 items below were directly inspected. Titles are descriptive audit labels. Audience and action recommendations are interpretations.

The older five-dish Reel is a positive counterexample to the generality of some recent captions. The community invitation gives explicit inclusions and a measurable keyword action. Event prices and dates in August posts are historical. Partner phone numbers belong to their organisers and must not be reused as SAZZ reservation contacts.

32 readable Instagram comment bodies were inspected across the 20 posts. They comprise 17 praise/community responses, four questions (three location requests and one social-media-team question), two event-keyword responses, two visit-experience reports, one companion-seeking comment, two tags, one organiser’s informational comment, and three unclear/off-topic responses. Categories are this audit’s interpretation. Counted but unreadable comments and unopened replies are excluded.

The clearest planning signal is three location questions on the creator collaboration. Two PUPPY keyword responses on the event invitation show interest in registration, not proven attendance or purchase. Two comments describe visits, one with a mixed/unclear reservation. An organiser comment repeats the historical price and registration action. No SAZZ reply was observed in the inspected bodies; unopened reply threads mean response quality remains unassessed. Creator visit · Event invitation

The ten District reviews were classified as seven positive, two mixed and one negative. Service, food and atmosphere recur; wait time, expense and taste appear as reservations. One reviewer describes a birthday arrangement. This is a small displayed sample, not representative satisfaction data. District reviews

Google Maps reviews and customer images

Yes: five full reviews and six reviewer-uploaded photos were inspected directly. Google Maps displayed 4.6 stars from 3,184 reviews. The initial five review cards were accessible; sorting or viewing more prompted sign-in, so newer and lower-rated samples could not be obtained. Their displayed ages ranged from three to six months. No owner reply was visible on these five cards.

The six photos, attached to three reviewers, show a pasta plate, two night views of the grounds, an entrance feature, seating with wood/woven textures, and a labelled chicken-katsu-ramen bowl. They visually support the warm, textured venue story and food variety. They do not verify current service policies, portion consistency or availability. The photo viewer dates them May–July 2026. Photographer credit and Google attribution are retained in the report’s screenshot excerpts.

Interpretation: Instagram’s inviting atmosphere broadly matches these customer-uploaded views. Add the practical questions that a group organiser needs answered. Do not infer a representative customer mix from five favourable, platform-selected reviews. Google Maps

The success metric is primary for each post, not a promised outcome. Reach-normalised metrics require supplied Insights. If unavailable, record the numerator and explicitly leave the denominator unknown; do not substitute views for unique reach.

Proposed rhythm: week 1 posts 01–03; week 2 posts 04–06; week 3 posts 07–09; week 4 posts 10–12. This is a sequence, not a scheduled calendar. If assets are delayed, move an approved concept forward rather than advertising unavailable dishes. There is no Onam promotion in this forward plan; the inspected August campaign is historical.

Purpose: help a first-time planner move from atmosphere to a qualified table enquiry. Persona Evening planner. Format: five static slides, proposed 1080 × 1350 canvas. Primary references: Preparing a warm welcome for staff/setting, Welcome to SAZZ for venue depth and Food and a place to stay for food/place pairing. Reference selection is like-for-like for venue storytelling; the factual end card is a proposed extension.

Caption draft:

Planning your first visit to SAZZ? Find us on Ukkadam–Selvapuram Bypass Road in Coimbatore. Swipe for the location and service hours, then choose what you would like to try.

For a table enquiry, share your preferred date, time and number of guests through [approved reservation route]. The team will confirm availability.

Save this for your next dinner plan.
#SazzKitchenLounge #CoimbatoreDining

Shot list: exterior/entrance; approach or approved landmark; corridor with seating; table at actual service scale; two current dishes; uncluttered venue texture. Capture people only with permission. Do not imply the venue will always be empty or a particular seat is guaranteed.

Production rules: use warm natural light or practical evening lighting with neutral food whites. Avoid strong orange skin/food shifts. Proposed margins: at least 72 px; heading around 64–76 px; essential body around 40–48 px, tested at phone size. These are new production starting points, not measured existing brand rules. Use approved brand fonts; if unavailable, label a neutral sans-serif as a temporary proof choice. Keep approved logo in one consistent upper-right location on slides 1–4 and on the final card, with no distortion and at least half its height as provisional clear space. Resolve logo clear space against brand guidelines when supplied.

Required before release: preferred reservation number/link, current map pin and landmark wording, current hours, dish availability, original photos, logo/font files, manager sign-off. Do not automatically choose one of the three observed phone numbers.

Measurement: count attributable qualified enquiries over seven days; log date/party size without unnecessary personal information. Where possible, have staff ask which guide brought the enquiry. Track confirmed reservations separately. Compare against an equal-length previous period with similar posting and promotion; attribution remains imperfect.

Purpose: make the drink range easier to compare while retaining its visual character. Persona Menu explorer. Format: five-slide static carousel, 1080 × 1350. References: Six drinks to discover ingredient/drink structure, Welcome to SAZZ drink/venue context and Food and a place to stay paired imagery. Three candidates from the September post are below; confirm availability or replace with currently approved drinks.

Caption draft:

Creamy and tropical, caramel with matcha, or crisp and refreshing?

Explore Coco Lopez, Burnt Caramel Matcha and Midnight Butterfly Fizz at SAZZ. Swipe to find the one that sounds like your next sip.

Save your pick and ask the team about current availability through [approved contact route].
#SazzKitchenLounge #CoimbatoreFoodies

This draft paraphrases sensory descriptions in Six drinks to discover. Do not add alcohol-free, caffeine-free, sugar-free, vegan or ingredient/allergen claims without verification. Do not state current prices unless supplied.

Shot list: one clean front/three-quarter image per glass at matching scale; ingredient or preparation detail for each; one pour; one three-drink table view. For a reuse-only version, extract approved high-resolution stills from originals, not the audit screenshots.

Layout and grading: retain the two-part ingredient/drink composition on slides 2–4, with equal widths and consistent crop logic. Proposed lower text band occupies roughly 22–28% of the canvas. Use a dark neutral sampled from approved artwork, not an asserted brand hex value. Name at 60–72 px, taste line at 42–48 px, no more than three short lines. Aim for at least 4.5:1 contrast for functional text and inspect at 360 px width. Logo remains upper-right, with approved proportions and sufficient clear space. Soft side light should reveal glass and liquid without blowing out white lettering. Preserve drink colour; do not oversaturate matcha or add ingredients for decoration.

Dependencies: current menu, three drink approvals, accurate taste descriptions, logo/font assets and footage usage rights. No availability approval means use placeholders in the internal proof rather than publishing product promises.

Measurement: saves/reach at day seven, with drink enquiries as a secondary signal. Compare to a similar informational carousel at equal age and promotion status. If only public metrics exist, record visible likes/comments separately and leave saves/reach unknown.

Purpose: make visible craft useful by naming the dish and one true preparation detail. Persona Menu explorer. Format: proposed 20-second vertical Reel, 1080 × 1920. References: A bowl with a little drama plating, Behind every bite whole-product view and Made for you process-to-result logic. These references were sampled visually; this new edit's timing is proposed, not copied from their pacing.

Select one current dish. The A bowl with a little drama bowl is a candidate, but use the chef-approved menu spelling rather than assuming its on-screen “Curry Laksha” label is the official name.

Caption draft:

Meet [approved dish name].

[One chef-confirmed sentence explaining the ingredient or technique shown.] Finished with [verified finishing ingredient], and served with [verified accompaniment, or remove this phrase].

Would this make your next order? Ask the team about current availability through [approved contact route].
#SazzKitchenLounge #CoimbatoreEats

Shot list: finished hero; whole raw/prepared ingredient; hands during technique; close process detail; final garnish/pour; whole dish; table context. Film only normal, approved food handling. Record clean kitchen sound where practical; use cleared music if needed. No audio from the inspected Reels has been evaluated or assumed licensed.

Visual rules: warm but neutral food colour, controlled highlights, soft directional light and modest background separation. Preserve one dominant subject per shot. Proposed minimum title size 64 px and body 48 px; keep essential text inside a central safe area, away from platform controls and bottom caption overlays. Check the current Instagram preview before export. Approved logo upper-right on cover/final frame; avoid covering hands or the food. Cover: “[approved dish name]” and small “From prep to plate”. Create a 4:5-safe central crop from the 9:16 master. Do not imitate private chef techniques that were not actually demonstrated.

Dependencies: chef/manager product approval, menu name, recipe facts for public use, shoot access, staff consent, available dish, approved brand assets and audio rights. If no specific factual detail can be verified, use an honest visual preparation sequence and remove the unsupported line.

Measurement: dish-specific enquiries at day seven; supporting retention/average watch time if supplied. Compare one named-dish Reel with a similar atmospheric Reel at equal age and promotion status. Record the creative differences and avoid claiming causality from a single pair.

Review all three concepts beside their references, confirm product proportions and colours, test phone-size text, verify factual fields and route consistency, then approve the production batch. This plan contains no scheduled or published posts. The agency's delivery capacity, fees and results are not assumed.

A four-week food-discovery and visit-planning content pilot is the strongest pitch. Proposed scope: 12 feed posts, comprising six carousels, five Reels and one static post; three posts weekly if assets and approvals allow. Pillars: choose what to try, understand the craft, plan a visit. Deliver approved caption/CTA copy, three reusable layout families, Reel covers, a contact/claim checklist and a measurement sheet. Reuse approved footage where suitable, with one planned shoot session if feasible. Do not promise a volume before confirming capacity, turnaround, asset rights and access.

This is support for an already-developed identity. Current agency relationships and internal capacity are unknown; the offer can complement an existing team. Fees, agency credentials and results are unspecified. Start by confirming contact routing and available dishes, then approve the three lead briefs. Compare results at equal post ages (for example seven days), splitting organic and promoted distribution. Track qualified enquiries and attributable confirmed bookings alongside saves/reach and video retention where Insights are supplied. A qualified enquiry includes a relevant date, party size or specific menu question; deduplicate repeat contacts. Public likes are not the business outcome.