DFI Intelligence Console · Hong Kong · 7-day rolling
Live · DFIQ sync 06:32 HKT
Thursday · 21 May 2026 · 06:34 HKT

Good morning, Ella.
What do you want to do today?

Twelve playbooks. Three modules. One Context Graph sitting above DFIQ. Three things changed this weekend — see the urgent chips, ask the system anything, or open one of the three playbooks below.

▴ Try asking
✦ The library · go deeper
9 playbooks · grouped by module · click any to open
Module 01 More from Brand Health
3 playbooks · Ask · Why · Signal
Module 02 More from Market Share
3 playbooks · Brief · Ask · Signal
Module 03 More from Consumer Voice
3 playbooks · Brief · Ask · Why
▴ The architecture
DFIQ aggregates the data. CalvinBall makes the decisions.
5M+ yuu members 90+ digital channels 120+ home visits indexed 7,400+ stores · 12 markets ~20K tokens/query · Context Graph
← Back to home
DFI Intelligence ✦ Ask DFI AI

Ask DFI AI anything — reasoning across all 12 playbooks.

A natural-language layer that synthesizes across yuu, POS, home-visit transcripts, and 90+ digital channels in under 8 seconds. Every answer cites its sources. The system will name what's structural, window-bound, or cyclical — and tell you when confidence is below 90%.

For: Ella + insights team Avg. response: 8.2 seconds Sources fused: up to 7 per question Reasoning across: 12 playbooks · 3 modules
▴ What this version of Ask can and can't do
Ask reasons across the live DFI data sitting above DFIQ. It synthesizes — it does not predict outside observed patterns. Where confidence is below 90%, Ask names it. Where Kantar and the internal view disagree, Ask surfaces the gap rather than picking a side. If Ask doesn't have the answer, it says so and routes you to the playbook that does.
▴ Try asking
✦ Ask DFI AI
DFI × CalvinBall
Module 01 · Brand Health · The CEO Monday Canvas
Hong Kong · 7-day rolling · Generated for Ella Chan
← Back to playbooks
Module 01 · Brand Health ◐ Brief

The CEO Monday Canvas. Five banners. One view.

Every Monday at 06:30 HKT, a fresh canvas. Five DFI banners, one frame, three things that changed this weekend pinned at the top. Each banner has a status, a one-line headline, a recommended action and a KPI. The CEO opens the link, reads the canvas in 90 seconds, and goes into EXCO informed. This replaces the monthly deck.

For: Ella Chan · CEO Trigger: Monday morning · pre-EXCO Window: Last 7 days Generation: Auto · 8 seconds Refreshes: Every 4 hours
Three things changed this weekend · 21 May 2026
The CEO Monday Canvas · 5 banners, one view
Refreshed 06:32 HKT · synth 8.2s · 7-day window
Sources fused for this canvas
yuu loyalty · 5.04M HK members · 7d home visit · April–May 2026 cohort, 24 visits POS sell-through · 1,401 HK stores · daily scan e-commerce · 90+ channels · session + basket data Xiaohongshu + IG · 8 KOLs tracked · 14d competitor signal · Watsons, Citysuper, Hanaro Mart DFIQ aggregation · pass-through, no write-back
▴ Recovery ceiling · what we don't claim
This canvas reflects 7-day rolling data on five HK banners. It is not predictive forecasting. Of the seven signals above, six are based on patterns with prior precedent we can show; one — the 7-Eleven late-night cohort — is a new behaviour with no 12-month baseline. We've flagged it as such. Any vendor showing you predictions without naming the structural-vs-cyclical decomposition is selling forecasts, not facts.
▸ Next playbook For deeper share analysis
The Brand Health canvas tells you what changed. Module 02 · Market Share tells you where DFI is winning or losing share, decomposed by cause and ranked by recovery feasibility.
← Back to playbooks
Module 01 · Brand Health ✦ Ask

Ask anything about banner health. In plain English.

A natural-language interrogation layer for the Brand Health module. Type a question — about a banner, a category, a moment, a comparison — and get a synthesized answer in under 8 seconds, with every source cited. Sana asked us for "market signals on the pulse, not monthly decks." Ask is the pulse.

For: Ella + insights team Avg. response: 8.2 seconds Sources fused: up to 7 per question Replaces: ~3-week manual loop
How is Mannings really doing this week?
▴ Synthesis · 4 sources fused · 8.4s

Mannings HK is having an exceptional week — but the picture is more specific than the topline suggests. Premium derma is up +18.4% on yuu in the last 7 days, concentrated in Wan Chai and Tsim Sha Tsui. The lift is unusually narrow: 73% of the category gain sits in just four SKUs — La Roche-Posay Cicaplast, Cetaphil Daily Cleanser, Eucerin Hyaluron-Filler, Vichy Mineral 89. Outside derma, the rest of the Mannings portfolio is on plan.

Three signals converged in the same window. One: home visit ethnography from the April–May cohort flags post-CNY skin renewal behaviour returning to last year's pattern within 9 days — a recurring annual cycle. Two: 8 Xiaohongshu KOLs posted Cetaphil routines in 14 days, driving +62% engagement on derma content. Three: Watsons is showing low-stock signals on 3 of the 4 lift-driving SKUs in their PR and e-commerce; supplier intel puts recovery at 10–14 days.

The window is real, the demand is verified, the competitor is unable to fulfil. Recommended: 14-day push on the four SKUs across Causeway Bay, TST, IFC + DFIQ Media activation to the 84,000-strong yuu high-frequency derma cohort. Expected lift on Mannings derma in HK for the window: +6 to +9 share points.

What's the risk if we don't act in 14 days?
▴ Synthesis · 3 sources fused · 6.8s

The window closes when Watsons recovers stock. Supplier-side intelligence (compiled from public procurement notices, distributor lead-times, and PR statements) puts Watsons recovery at 10 to 14 days with ~80% confidence. Two scenarios:

If we act inside 14 days: Mannings captures the natural demand peak + the displaced Watsons demand + the KOL-driven inquiry surge. Projected lift: +6 to +9 share points in HK derma for the period. Margin impact: net positive across all four SKUs at current Mannings shelf price.

If we don't act: Mannings still benefits from the natural cycle (~+8 to +11% on the 4 SKUs, baseline), but the displaced Watsons demand reverts when stock returns. We forgo the asymmetric window. The cost isn't dollars — it's a missed 14 days that won't repeat until the next post-event cycle.

▴ What Ask cannot do · in this version
Ask synthesizes existing signals — it does not make predictions outside documented behavioural patterns. The "10–14 day Watsons recovery" estimate above is from public supplier signals with ~80% confidence, not an internal data point. Where confidence is below 90%, Ask shows it. When in doubt, Ask says so rather than inventing an answer.
▴ Brilliant questions for this view
✦ Ask Mode
← Back to playbooks
Module 01 · Brand Health ◎ Why

Why did Mannings derma surge this week? Three signals, one window.

When a banner moves, ask why. The Why module decomposes the health movement into ranked contributing causes — each backed by a different data source and confidence score. Honest about what's cyclical, structural and tactical, so leadership knows which causes to invest behind and which to ride.

Target metric: Mannings derma · +18.4% on yuu · 7d Causes ranked: 3 of 5 contributing Confidence: High on all 3
Contribution waterfall · +18.4pts to be explained
+24% +16% +8% +4% 0 0 BASELINE +8.1 POST-CNY CYCLE · 44% +5.5 WATSONS STOCKOUT · 30% +4.8 KOL SURGE XIAOHONGSHU · 26% +18.4% 7-DAY ACTUAL
1
Post-CNY skin renewal cycle Recurring annual pattern · home visit cohort confirms behaviour returning within 9 days of 2025 baseline · 11 of 16 respondents in catchment · CYCLICAL · structural to category
+8.1pts44% contribution
2
Watsons stockout · 3 of 4 lift-driving SKUs Public PR + e-comm signals · low-stock on La Roche-Posay, Cetaphil, Eucerin · supplier recovery 10–14d (~80% confidence) · TACTICAL · window-bound
+5.5pts30% contribution
3
KOL surge on Xiaohongshu + IG 8 KOLs in 14d posting Cetaphil routines · +62% engagement on derma content vs trailing month · SHORT-CYCLE · 30–60d lift
+4.8pts26% contribution
Sources yuu loyalty · derma category, 7d home visit · April–May cohort, 24 visits Mannings POS · 14 stores HK Xiaohongshu + IG · 8 KOLs Watsons stock + PR signal
▴ What's structural vs. what's tactical · read this before the EXCO
Of the +18.4 point surge, ~+8.1 is cyclical — Mannings will see this every post-CNY regardless of what we do. +5.5 is window-bound on Watsons stockout, which closes in 14 days. +4.8 is short-cycle social driven by KOL momentum that typically decays in 30–60 days. The opportunity to act is on the +5.5 window-bound piece — it's the only one our action accelerates. Any pitch claiming we "captured +18.4pts" is taking credit for a cycle we didn't drive.
▸ Recommended action For Wednesday EXCO · 27 May
Capture the +5.5pt window. 14-day push on the 4 SKUs + DFIQ Media activation to 84K yuu derma cohort. Cost: existing Mannings Q2 trade spend reallocation, no incremental. Expected category lift: +6 to +9pts.
← Back to playbooks
Module 01 · Brand Health ◭ Signal

Forward radar. What's coming before it shows up in the monthly deck.

Anomaly detection across yuu loyalty, sell-through, home visit and digital channels — surfacing emerging banner-health shifts before they hit the monthly read. Configurable thresholds, severity-ranked, with the trigger conditions documented per signal.

Active signals: 7 Watch list: 12 items Avg lead time: 14 days Refresh: Every 4 hours
7 active signals · today
38 min ago High · Window
Mannings premium derma · +18.4% on yuu, 73% in 4 SKUs · Watsons in stockout · 14-day window Triggered by: yuu basket lift threshold (+15% / 7d) + competitor stockout signal · sources: yuu, home visit, social, competitor PR
6h ago High · Defence
Wellcome Central · share −2.3pts in 7d · Hanaro Mart proximity · yuu defection rate 12.4% (baseline 3.1%) Triggered by: share-loss threshold (−2pts / 7d) + new-competitor catchment detector · sources: yuu, POS, geo-catchment
12h ago Medium · New cohort
7-Eleven late-night · basket +14% YoY for 11pm–2am window · new segment confirmed Triggered by: time-of-day anomaly across 632 stores + repeat-visit signal · confidence: medium · single CNY 2026 datapoint, no 12-month baseline
14h ago Medium · Macro
IKEA Kowloon Bay · foot traffic −9.2% · HKD strength impact on cross-border Shenzhen day-trippers Triggered by: traffic threshold (−7% / 14d) + FX correlation flag · sources: footfall + competitor benchmark + macro feed
22h ago Medium · Ceiling
Maxim's reunion sets · CNY 2026 price elasticity ceiling reached · volume flat at +8% price Triggered by: price-volume decoupling pattern · confidence: high · pattern matches 2024 Lai See observation
2d ago Watch · Healthy
Guardian vitamin D · +22% on yuu in 45–60 segment · aligned with HK DoH campaign · sustainable Triggered by: demographic-overlay anomaly + sustained 14-day signal · classification: opportunity to expand
3d ago Watch · Premium
Market Place organic produce · +11% in Mid-Levels & Repulse Bay · premium customer base growing Triggered by: catchment-level basket growth · classification: healthy growth, Citysuper benchmark holding
90-day forward radar · all 5 banners
Timeline · what's emerging in the next 90 days 7 active · 12 watch · 3 high-priority events ahead
TODAY · 21 MAY May 21 Jun 20 Jul 20 Aug 19 Mannings 14d window +6 TO +9 PTS Wellcome 30d recovery 11/11 prep window +30 DAYS Guardian VitD range 11/11 launch 90 DAYS · LIVE Black Friday prep
High priority
Medium · monitor
Planning window
Watch only
▴ What this radar doesn't predict
Signal is an anomaly engine, not a forecast. It detects unusual patterns in observed data and flags them for human review with confidence scores. It does not predict price changes, competitor moves Hanaro hasn't announced, or category shifts without precedent. The 14-day Watsons recovery estimate is a public-signal extrapolation, not an internal supply-chain feed. When confidence is below 90%, we say so.
DFI × CalvinBall
Module 02 · Market Share · Group Share Heatmap
Hong Kong · 30-day delta · Generated for Ella Chan
← Back to playbooks
Module 02 · Market Share ◐ Brief

Group Share Heatmap. Banner × category × geography.

The view Ella shows the chairman. Live share movement across every banner, every category, every HK geography — color-coded by 30-day delta, prescriptive flags where action is overdue. 5 banners × 6 categories × 3 HK geographies = 90 cells, refreshed every 4 hours.

For: Ella Chan + group strategy Cells live: 90 Refresh: 4h Window: 30-day rolling delta
HK share movement · 30-day delta · 21 May 2026
Banner Food & Beverage Health & Beauty Home & Living Premium / Imported Value / Own-Label Convenience
Wellcome HK · Central −0.1 N/A N/A +0.3 −2.1 N/A
Wellcome HK · Mid-Levels +0.2 N/A N/A +0.8 −0.2 N/A
Wellcome HK · Kowloon +0.1 N/A N/A +0.2 −0.4 N/A
Mannings HK · Central N/A +3.6 N/A +4.2 +0.1 N/A
Mannings HK · Kowloon N/A +1.8 N/A +2.4 −0.1 N/A
7-Eleven HK · all +1.1 +0.3 N/A +1.6 +0.4 +2.8
Maxim's HK · all −0.3 N/A N/A −1.1 −0.2 N/A
Guardian HK · all N/A +1.4 N/A +0.9 +0.2 N/A
IKEA HK franchise N/A N/A −1.8 N/A N/A N/A
Where attention should go this week
Mannings Central · Premium Strongest gain
+4.2pts

Derma surge + competitor stockout. Window-bound 14d. Action: 14-day premium SKU push (Module 01 · Brief).

Wellcome Central · Value Largest loss
−2.1pts

Hanaro Mart catchment loss · price-sensitive segment defecting. Action: localised Korean SKU + yuu push (Module 02 · Why).

7-Eleven · Convenience New cohort win
+2.8pts

Late-night basket emergence · CNY 2026 reopening signal. Action: lock 2027 playbook (Module 03 · Brief).

▴ Limits of this heatmap
90 cells, refreshed every 4 hours, on a 30-day delta. This is directional, not absolute share — DFI internal yuu signal + POS + a competitor-overlay model with documented confidence per cell. For absolute share by category we still defer to Kantar / Nielsen. Where Kantar disagrees with this view, we show both — and the gap is itself a signal.
← Back to playbooks
Module 02 · Market Share ✦ Ask

Where am I losing share? Ask in plain English.

Natural-language interrogation of the share data — across banner, category, geography, segment. Ella's questions, in her language, answered with cited evidence and a recommended lever. Built on the same Context Graph that powers the heatmap — every answer reconciles back to a cell you can audit.

For: Ella + category leads Avg. response: 7.8 seconds Audit trail: every cell traceable
Where am I losing share in HK Central derma?
▴ Synthesis · 5 sources fused · 7.4s

You're not. In HK Central derma you're gaining — Mannings Central is up +3.6pts in Health & Beauty over 30 days, and +4.2pts in Premium / Imported. The gain is driven by the post-CNY skin renewal cycle + the Watsons stockout window flagged in Module 01.

The Central catchment is losing share — but not in derma. Wellcome Central is down −2.1pts in Value / Own-Label, driven by Hanaro Mart opening at 28 Queen's Road Central on 28 April. The bleed is value-tier only; premium tier is intact at +0.3pts.

If your question is really "where am I exposed in HK Central this quarter," the answer is Wellcome Value, not Mannings Derma. Want me to open the Wellcome decomposition?

Which 3 cells in the heatmap need attention this week?
▴ Synthesis · 4 sources fused · 6.1s

Ranked by action-urgency, not absolute magnitude:

1 · Mannings Central Premium · +4.2pts — window-bound, 14 days. If you don't act by Wednesday the +5.5pt window-piece reverts when Watsons restocks. Highest urgency.

2 · Wellcome Central Value · −2.1pts — recoverable +1.3pts in 30 days on existing budget if Korean SKU range + yuu reactivation goes live. Highest dollar.

3 · 7-Eleven Convenience · +2.8pts — not urgent this week, but the 2027 playbook decision needs to be locked before Q3 to plan supply and assortment. Highest strategic.

The other 87 cells are within their normal-variation envelope — flat to ±0.5pts. No EXCO attention required.

Compare DFI total HK share movement to Kantar's last read
▴ Synthesis · 3 sources fused · 5.9s

DFI internal signal (this view, 30-day delta, May 2026): +0.4pts net across all HK banners, with category-level variance from −2.1 (Wellcome Value Central) to +4.2 (Mannings Premium Central).

Kantar last read (Q1 2026 wrap, published 14 May): +0.2pts net for DFI HK. Directionally aligned, magnitudes differ by ~0.2pts at the topline.

The gap is itself a signal. Kantar lags by ~45 days, samples 4.2K HK households, and weights by demographic — it underweights Central / Mid-Levels affluence by an estimated 8%. Our internal view is faster but skews toward heavy-yuu users; Kantar is slower but broader. Triangulate before any strategic call. For Mannings derma the gap is small enough to act on; for Wellcome the two views agree the bleed is real.

▴ What Ask does & doesn't see · in this module
Ask reconciles to internal signal: yuu, POS, geo-catchment, competitor overlay model. It does NOT have direct read access to Kantar or Nielsen — third-party panel data is referenced as a fact-checked comparison, not synthesized. Any answer that crosses the third-party threshold cites the comparison source explicitly. If you need a Kantar-grade answer, Ask says so and stops.
▴ Brilliant questions for this view
✦ Ask Mode
← Back to playbooks
Module 02 · Market Share ◎ Why

Why is Wellcome Central losing share? Three causes. One recoverable.

Wellcome Central lost −2.3 share points in 7 days. The drop is not brand-wide. It's a localised value-tier bleed with one structural cause, one decay pattern and one solvable trigger. Each cause is decomposed, sourced, and rated for recovery feasibility — so leadership can fund where it matters and accept what it can't reverse.

Target metric: Wellcome Central · −2.3pts · 7d Causes ranked: 3 of 4 contributing Recoverable: +1.3pts in 30d Confidence: High on causes 1 + 2 · Medium on 3
Contribution waterfall · −2.3pts decomposed
0 −0.5 −1.0 −2.0 −3.0 0 BASELINE −1.0pt HANARO PROXIMITY · 44% −0.7pt VALUE-TIER ELASTICITY · 29% −0.6pt YUU ENGAGE DECAY · 27% −2.3pts 7-DAY ACTUAL
1
Hanaro Mart proximity opening · 28 Apr New Korean grocery at 28 Queen's Road Central, 400m from 3 Wellcome stores · 47.2K affected yuu members in catchment · defection rate 12.4% vs baseline 3.1% · PARTIALLY RECOVERABLE · ~50% recapture realistic
−1.0pt44% contribution
2
Value-tier elasticity in Central catchment Price-sensitive segment more elastic to alternatives · Own-label rice −8.4%, fresh produce value-tier −6.2% · premium tier flat at +0.3% · STRUCTURAL · accept, don't fight
−0.7pt29% contribution
3
yuu engagement decay · pre-existing Engagement softening in Wellcome Central yuu cohort began 60 days before Hanaro opened · reduced personalised offer frequency · FULLY RECOVERABLE · reactivation programme
−0.6pt27% contribution
Sources yuu loyalty · 47.2K Central members Wellcome POS · 3 affected stores · 30d home visit · Hanaro shopper interviews · 11 sessions Hanaro Mart launch · public + foot traffic Wellcome.com.hk · session decay · 60d
▴ Recovery ceiling · what's recoverable vs structural
Of the −2.3pts, we project +1.3pts recoverable in 30 days on existing budget: +0.5pts from Hanaro recapture via Korean SKU range + price-matching on 14 lines + targeted yuu offers, and +0.6pts full recovery on yuu engagement decay via reactivation campaign. The remaining −0.5pt from Hanaro is permanent customer loss (the price-sensitive segment that prefers Hanaro's range) and the −0.7pt value-tier elasticity is structural — chasing it down with promotion would erode margin without holding the customer. We do not recommend it.
▸ Recommended actions For Wellcome category review · 28 May
Two moves, existing budget. (1) Korean SKU range expansion at 3 Central stores + price-match on 14 lines vs Hanaro. (2) yuu reactivation push to 47.2K affected members + personalised offer cadence reset. Projected recovery: +1.3pts in 30d.
← Back to playbooks
Module 02 · Market Share ◭ Signal

Catchments at risk. Find the next Wellcome Central before it bleeds.

A forward radar for share defence. Watches competitor leases, store openings, planning applications, yuu defection precursors and segment-level basket softening — flagging at-risk catchments 30 to 90 days before share visibly moves. The view that turns Wellcome Central from a fire drill into a project plan.

At risk: 4 catchments Watch: 9 catchments Avg lead time: 42 days Refresh: Daily · 06:00 HKT
4 catchments at active risk · ranked by exposure
+0 days High · Live
Wellcome Central · Hanaro Mart impact −2.3pts share · 47.2K yuu members in active defection Detected: 02 May · Action: localised range + reactivation programme · See Module 02 · Why for decomposition · Projected recovery +1.3pts / 30d
−18 days High · Building
Wellcome Tai Po · Don Don Donki opening 8 Jun · 600m proximity to flagship + 2 satellite stores · ~28K yuu members in catchment Detected via: planning permit + fit-out activity + competitor PR · Hanaro precedent suggests 1.5–2pt share impact if no pre-defence · pre-launch lead time 18 days remaining
−42 days Medium · Watch
Wellcome Shatin · yuu engagement decay pattern matches pre-Hanaro Central signal · session frequency −9% in 60d Detected: yuu engagement decay model · no competitor signal yet, structural attention required · same root cause as Wellcome Central yuu decay layer (cause 3, Module 02 · Why)
−65 days Medium · Building
Mannings Kwun Tong · Watsons relocating flagship within 200m · scheduled ~25 Jul 2026 Detected via: lease records + Watsons internal signal · current Mannings Kwun Tong basket +6.4% · counter-positioning window: 50 days · category mix decision needed
9 catchments on watch · no active intervention
Wellcome Mid-Levels42d watch

Premium tier intact · monitoring for FX-led Shenzhen substitution behaviour

Mannings TST28d watch

Watsons promotional cadence signal · 14-day trial run flagged · no impact yet

IKEA Kowloon Bay14d watch

HKD strength macro · cross-border behaviour soft · monitor for stabilisation

7-Eleven Causeway Bay35d watch

Late-night cohort emergence · classification as opportunity, not risk · monitor

Maxim's Festival Walk21d watch

Premium reunion tier pricing ceiling effect · monitor for Q4 mid-autumn signal

Guardian Kowloon City19d watch

VitD category sustainable · Watsons promotional response watch · low concern

Wellcome Lai Chi Kok45d watch

Park N Shop new flagship 1.2km · monitoring proximity threshold breach

7-Eleven Sai Ying Pun33d watch

FamilyMart pilot expansion signal · no proximity threats yet · early radar

Mannings IFC25d watch

Sephora HK premium expansion signal · derma overlap risk · low impact today

▴ What the radar can & can't see
Detection inputs: lease records, planning permits, public competitor PR, fit-out activity, yuu engagement decay model, basket softness anomalies. Confidence is highest for documented physical openings (Hanaro, Don Don Donki, Watsons relocation) and lower for pricing or assortment moves by competitors which are private until they ship. The radar will not detect a competitor's pricing strategy change until it shows up in our POS or yuu defection data — usually a 14-day lag.
DFI × CalvinBall
Module 03 · Consumer Voice · Event Playbook
Hong Kong · CNY 2026 wrap + 11/11 prep · Generated for Ella Chan
← Back to playbooks
Module 03 · Consumer Voice ◐ Brief

Event Playbook · live. 11/11, Black Friday, CNY — pre / during / post.

DFI's biggest commercial moments turned into live playbooks. CNY 2026 wrap is done. 11/11 prep window is open. Black Friday is staged. CNY 2027 lessons are loading. Each event has its own pre / during / post discipline — built from home visit, loyalty, sell-through and social fused in real time.

Live events: 4 in rotation Wrapped: CNY 2026 · 11 days post Next: 11/11 2026 · 174 days One event ROI: pays for the platform
Wrapped · 10 Feb 2026 CNY 2026 ● Post · synthesis complete
+174 days · 11 Nov 2026 11/11 2026 ◆ Pre · prep window open
+186 days · 23 Nov 2026 Black Friday 2026 ◆ Pre · staged
+265 days · 10 Feb 2027 CNY 2027 ○ Future · loading 2026 lessons
▴ CNY 2026 · post-event synthesis
Three formats. Three different stories from one CNY.
7-Eleven · NEW COHORT+14% basket YoY

Late-night CNY behaviour emerged for the first time in 2026 — driven by Causeway Bay nightlife reopening + post-event single-serve premium snacking. Lock the 2027 playbook now.

Decision: lock late-night assortment + supply for CNY 2027 · pilot for 11/11 2026 (12 flagship)
Maxim's · CEILINGVol flat at +8% price

Reunion sets hit pricing ceiling — premium tier −6%, value tier held. The 5-year price-up strategy has reached elasticity. Repricing study needed for CNY 2027.

Decision: commission price-volume elasticity study by Mar 2027 · do not raise reunion prices further
Wellcome · MISSEDGifting window −14d

CNY gifting peak shifted 14 days earlier in 2026 vs 2024 baseline (driven by office return calendar). Wellcome gifting display caught wave too late. Lost 4–6pts of share to Citysuper.

Decision: shift CNY 2027 gifting display calendar forward 14 days · brief Wellcome merchants by 1 Sep
CNY 2026 sources home visit · 18 households · gifting + reunion ethnography yuu basket · 5.04M HK members · CNY ±10d POS · all banners · time-of-day disaggregation Xiaohongshu + IG · 47 KOLs · CNY content tracked Citysuper / ParknShop benchmark · gifting window timing
▴ 11/11 prep · 174 days outreadiness: 38% · CNY 2026 lessons loading into playbook
Late-night assortment
12-store pilot from CNY '26 cohort → 11/11 test launch
Gifting calendar
14-day earlier launch for Wellcome gifting · approved Jul
DFIQ Media activation
3 cohorts targeted · pending Wee-Lee security review
Mannings derma
Sustainable +22% VitD trend extended into 11/11 health push
▴ One-event ROI · how it pays for the platform
The CNY 2026 Wellcome gifting miss alone cost an estimated 4–6 share points in HK gifting category across 3 weeks. Even at the lower bound, the recovery for CNY 2027 from a single calendar shift more than covers an annual platform subscription. One event win pays for the platform. The other 3 events are pure margin.
← Back to playbooks
Module 03 · Consumer Voice ✦ Ask

Ask the ethnography directly. 120+ home visits, instantly queryable.

DFI commissions home visit research every quarter — the richest signal you have, almost never queried. Now it is. Type a question about consumer behaviour, gifting decisions, format choice, brand affinity — and get a synthesized answer with transcript fragments cited verbatim and behavioural patterns flagged across visits.

Indexed: 120+ home visits · 2024–2026 Languages: Cantonese / English / Mandarin Avg response: 9.6 seconds Citation: transcript line + visit ID
What did home visits tell us about CNY 2026 gifting decisions?
▴ Synthesis · 18 visits · 9.2s

Three patterns emerged across 18 CNY-window home visits (24 Jan – 14 Feb 2026):

1 · The gifting window shifted 14 days earlier than 2024. 11 of 18 households began gifting purchases before the start of the official CNY shopping window because the post-pandemic office return calendar pulled in-person gifting forward. "I had to give Lai See to my team at the office party on the 28th — I went to Citysuper on the way home that Friday." [Visit HV-2026-014, female, 41, Wan Chai].

2 · Wellcome was not the first stop for premium gifting. Citysuper and Marketplace by Jasons took the early window. Wellcome was the second-choice ("if Citysuper was out") — 3 of 18 respondents named Wellcome as their primary gifting destination, vs 11 for Citysuper. "Wellcome is for groceries — when I'm giving a gift I want something more special." [HV-2026-009, female, 36, Causeway Bay].

3 · Premium pricing ceiling at Maxim's reunion sets is visible in the conversation. 6 of 12 reunion-set buyers spontaneously commented on price. "They've gone up every year. This year I downgraded — got the Bronze instead of Gold." [HV-2026-005, female, 52, Kowloon Tong]. The 5-year price-up has reached the ceiling Maxim's POS confirms.

The synthesis maps cleanly to the Module 03 · Brief findings, with the ethnography providing why in the consumers' own words.

Are 7-Eleven late-night customers genuinely new, or returning?
▴ Synthesis · 9 visits · POS overlay · 7.8s

Genuinely new. 5 of 9 late-night 7-Eleven visitors interviewed in March 2026 reported they had not used 7-Eleven for late-night in the prior 18 months. The pattern is post-pandemic re-emergence rather than first-time discovery: "I used to go before Covid, then I stopped. Now Causeway Bay's open again and I'm walking past 7-Eleven at 1am." [HV-2026-031, male, 28, Wan Chai].

yuu basket data triangulates: 62% of the late-night transaction surge comes from members whose 12-month prior late-night frequency was zero. This is a returning cohort behaviourally — a customer base coming back, not new acquisition.

Implication for CNY 2027: the late-night assortment decision shouldn't be designed for occasional visitors — it should be designed for habits that are reforming. Premium single-serve + convenience meals + impulse alcohol. Not snack-first.

▴ Ethnography is signal, not statistics
120+ visits is rich but is not a statistically representative sample. Patterns flagged from ethnography are hypothesis-generating; they get cross-referenced against yuu, POS or social to confirm scale. Every Ask response in this module says how many visits the answer is based on, and where the pattern has been quantitatively triangulated vs left as qualitative-only. If a finding hasn't been triangulated, we say so.
▴ Brilliant questions for this view
✦ Ask Mode
← Back to playbooks
Module 03 · Consumer Voice ◎ Why

Why did the Maxim's reunion set hit a pricing ceiling? Three forces. One consumer.

Maxim's CNY 2026 reunion sets held volume flat at an 8% price increase. The 5-year price-up strategy has hit elasticity. This Why surface decomposes the consumer-side cause across ethnographic evidence (what they said), behavioural evidence (what they did), and competitor benchmark (what they did instead). One synthesized answer with three independent sources — the moat Sana asked for.

Target: Maxim's reunion · volume flat at +8% price Sources fused: 3 Confidence: High
Three-source convergence · all three say the same thing
▴ Home visit ethnography12 visits

6 of 12 reunion-set buyers spontaneously mentioned price. Premium-tier downgrading behaviour was named in 4 separate interviews.

"They've gone up every year. This year I downgraded — got the Bronze instead of Gold."
— HV-2026-005 · female, 52 · Kowloon Tong
▴ POS behavioural data280 Maxim's stores · CNY ±10d

Premium-tier reunion volume −6% · Bronze-tier reunion volume +9%. Customer is not leaving — they are downgrading within the brand.

Same 5-year customer cohort showing tier substitution, not category exit. Total reunion volume held; mix shifted down one notch. Margin per set down ~12% as a consequence.
▴ Competitor benchmarkCitysuper · Imperial Treasure

Citysuper held reunion price flat YoY · Imperial Treasure raised +4%. Maxim's at +8% became the ceiling outlier in the category.

Switching cost is low — most buyers consider 2–3 options. The relative price gap (Maxim's now 14% premium over Citysuper, was 6% in 2024) is what changed, not absolute pricing.
▴ Synthesized answer · what all three sources agree on

Maxim's reunion sets have not hit a price ceiling because consumers can't afford them. They have hit a ceiling because the relative price gap to Citysuper widened from 6% to 14% over five years while perceived quality stayed comparable. Faced with a 14% premium, the existing Maxim's customer is choosing to stay with the brand but downgrade the tier — captured in ethnography ("I got the Bronze instead of Gold") and confirmed in POS (premium −6%, Bronze +9%).

The strategy implication is not "lower the price" — it is "don't widen the gap further." Margin per set has already dropped ~12% from tier substitution. Another 4% price-up in 2027 would risk crossing the brand-switch threshold (estimated 18% relative premium based on Citysuper customer share-of-wallet data) at which point downgrading is replaced by leaving.

▴ What this Why doesn't model
This decomposition tells you why Maxim's lost margin at CNY 2026 — it does not predict the consumer-switch threshold beyond Citysuper. If Imperial Treasure repositions, or a new premium player enters, the threshold shifts. Run the repricing study Maxim's needs by March 2027 — this is not a forecast we can substitute for that work.
▸ Recommended action For Maxim's leadership · 2 Jun
Hold reunion pricing flat for CNY 2027 · commission elasticity study by March 2027 · run a Bronze-tier upgrade experiment in Q4 to test if perceived-quality investment can recover premium-tier volume without crossing the switch threshold.
← Back to playbooks
Module 03 · Consumer Voice ◭ Signal

Event radar. Every retail moment, 90 days forward.

A live calendar of DFI's commercial moments — what to lock now, what to test, what to retire from last year's playbook. Each event has lead-time alerts, prep-window thresholds and a confidence score based on prior-year wrap. Today: 11/11 prep window opens in 14 days, Black Friday is staged behind it, CNY 2027 lessons are loading.

Events tracked: 11 retail moments Open prep windows: 3 Decisions overdue: 0 Avg lead-time captured: 72 days
90-day event radar · ahead of every retail moment
TODAY · 21 MAY May 21 Jun 20 Jul 20 Aug 19 PREP OPEN 11/11 pilot lock +14 DAYS Mid-Autumn prep Q3 VitD push 11/11 LIVE +174 DAYS Black Friday +186 DAYS
Prep window open
Pilot / pre-launch lock
Calendar build
Event live
Sequenced behind
Per-event decisions to lock · across 11/11, Black Friday, CNY 2027
+14 days 11/11 pilot
Lock 7-Eleven late-night assortment · pilot 12 Causeway Bay flagship stores by 4 Jun · pilot data feeds 11/11 launch decision Decision owner: 7-Eleven category director · Inputs ready: CNY 2026 wrap, ethnography (Module 03 · Ask) · Approval lead time required: 5 days
+30 days 11/11 pre-launch
Wellcome gifting calendar — earlier launch lock · 14-day shift forward · brief merchants by 21 Jun Source decision: CNY 2026 missed-window finding (Module 03 · Brief) · Owner: Wellcome merch director · Re-test: gifting display from 25 Oct vs 8 Nov
+45 days 11/11 + BF media
DFIQ Media activation plan · 3 cohorts × 5 banners · pending Wee-Lee security & permissions review Currently blocked on DFIQ data-pull permissions sign-off · Wee-Lee escalation needed by 5 Jul · DFIQ stays system of record, CalvinBall does not write back
+186 days Black Friday
Black Friday 2026 staged · positioning brief due 60 days before launch · sequenced behind 11/11 with shared assortment for premium tier No decisions overdue · current readiness 22% · prep window opens 8 Sep
+265 days CNY 2027
CNY 2027 lessons loading from 2026 wrap · 3 streams locked: 7-Eleven late-night, Wellcome gifting +14d, Maxim's no-further-price-rise Prep window opens 20 Sep · Maxim's repricing study commission deadline: 1 Mar 2027 · readiness 18%
▴ What the event radar doesn't do
This radar tracks retail moments DFI can plan around — 11/11, Black Friday, Mid-Autumn, CNY, Mother's Day, Father's Day, Lai See, Mooncake, Easter, Christmas, school-return. It does not anticipate black-swan events (typhoons, geopolitical shocks, supply disruptions, public health). Those require human intervention against this calendar — the radar is the platform on which leadership reacts, not a substitute for leadership.
▴ Live update