A 30/60/90 plan. Month one builds the AI layer. Month two builds the campaign on the layer. Month three hardens what's running.
In short: If I had 90 days until 9.9, here's the plan I'd run. Month one is not for finalising the brief — it's for building infrastructure: marketing intelligence on last year's data, conversation commerce in WhatsApp and platform chat, integrated D2C plumbing across surfaces. Month two briefs the agency against those findings, specs the creative for variant volume, wires every ad into chat. Month three pre-warms the platforms, runs conversation commerce dress rehearsals, locks daily marketing intelligence reads, builds in-thread cart recovery. Decide the parallel-team question in week one. Everything else fits the 90 days only if that decision is made early.
Contents
Most marketing heads with 90 days until 9.9 are spending those days in the wrong order. The first month goes on refining the campaign idea — brief iterations, agency presentations, budget debates. The second month goes on creative production. The third month is the campaign.
That sequencing made sense in 2024. In 2026, it produces a 9.9 plan that sits on no infrastructure. The marketing intelligence findings that should have shaped the brief arrive after the brief is locked. The conversation commerce layer that should have been trained in month one gets bolted on in week eight. The integrated D2C plumbing that attribution depends on is still being tested when the campaign goes live. Here's the 30/60/90 plan I'd run instead.
Month one is not for finalising the brief — it's for building infrastructure. Marketing intelligence on last year's 9.9, 11.11, and 12.12 data. Conversation commerce coverage trained, wired, and tested. Integrated D2C plumbing across Shopee, Lazada, brand site, and WhatsApp. Without this layer, everything in month two is guesswork.
The 30-day infrastructure sprint has three tracks running in parallel.
Marketing intelligence on last year's data. AI applied to first-party data — Shopify events, customer profiles, platform purchase history, post-campaign surveys from the last three sale windows. The output is the brief-shaping answers: which segments responded to which pricing mechanics, which SKUs had high pre-purchase intent signals but underperformed on 9.9 day, which cohort from last 9.9 has the highest 12-month LTV. These answers shouldn't arrive after the brief is written. They should be in the room when the brief gets written. That's only possible if the work starts in week one.
Conversation commerce coverage trained and wired. The AI agents that will handle pre-cart questions across WhatsApp and Shopee and Lazada chat during 9.9 need to be trained on the product catalogue, brand voice, pricing rules, bundle mechanics, and the most common pre-purchase questions from last year's campaign. Escalation logic between AI and human needs to be defined and wired. In-thread recovery flows need to be built. This is not a two-week job. It's a 30-day job at minimum. Starting it in month two is too late.
Integrated D2C plumbing across surfaces. The integration layer that connects Shopee, Lazada, brand site, WhatsApp, and email into a single customer view needs to exist before the campaign goes live — not as a post-campaign analytics project. Without it, chat-surface conversions don't attribute correctly. Retargeting runs blind. Recovery flows don't know what the customer already did on another surface. Month one is when this gets built and tested under non-live conditions.
The brief gets written at the end of month one — not at the start. By day 30, the marketing intelligence findings are in the room, the conversation commerce layer is built, and the integrated D2C plumbing is tested. The agency brief that comes out of that room is a structurally different document from one written on day one with last year's assumptions.
Brief the agency with marketing intelligence findings, not last year's personas. Spec creative for variant volume — 30 to 50 assets — not hero polish. Wire every ad into the conversation commerce layer. Pressure-test integrated D2C plumbing under load. The campaign sits on the infrastructure, not next to it.
Month two is the conventional campaign-build month — brief, creative production, media planning, audience setup. The difference is that in 2026, each of those workstreams is running on the infrastructure built in month one, not on assumptions.
The agency brief comes from the marketing intelligence findings. The brief tells the agency which segments to prioritise based on actual LTV and repurchase data, which SKUs to push based on intent signals from last year's pre-9.9 window, which creative angles performed against which audience cohorts. The agency is building against real first-party insight, not last year's brand personas.
Creative is specced for variant volume, not hero polish. The brief specifies 30 to 50 creative variants — different hooks, different product angles, different audience framings — alongside the brand guidelines and tone-of-voice rules the AI creative pipeline needs to operate within. The agency owns the strategy. AI generates the volume. The platform tests and scales the winners. Hero assets are still produced, but they're not the primary bet.
Every ad creative is wired into the conversation commerce layer. Each ad has a clear path — product page or chat. Where the expected pre-purchase question load is high, the ad routes into WhatsApp or platform chat directly, not into a product page the buyer has to search for answers on. This is the brief change that requires conversation commerce to be built before month two — the creative can't be designed for it if the layer doesn't exist yet.
Integrated D2C plumbing gets pressure-tested. Month two is when the integration layer gets load-tested — simulated traffic across surfaces, attribution verified across channels, recovery flows tested end-to-end. Any gaps in the integrated D2C layer discovered now are 10x cheaper to fix than gaps discovered during a live 9.9 campaign.
By day 60, the creative pipeline is running, the media plan is briefed, the conversation commerce layer is trained and wired into the creative, and the integrated D2C plumbing has been tested under simulated load. Month three is operationalisation — not build.
Pre-warm platforms with small spend two weeks out. Run a live conversation commerce dress rehearsal through ad → chat → checkout. Lock daily marketing intelligence reads for the campaign window. Build in-thread cart abandonment recovery. Operationalise — don't add more — in the last 30 days.
Month three is not for building new things. It's for ensuring everything built in month one and two is operational under live conditions. Four workstreams.
Platform pre-warming. Running small spend through Shopee's advertising engine, Lazada's sponsored listings, and Meta's campaign infrastructure in the two weeks before 9.9 gives the platform AI systems signal to work with before the campaign window opens. Cold accounts on day one of 9.9 get reallocated against — the platform AI has no basis to favour them. Pre-warmed accounts have purchase signals, engagement data, and algorithm familiarity. Day one of 9.9 starts warm, not cold.
Conversation commerce dress rehearsal. Running a live end-to-end test of the full conversation commerce flow — ad → chat surface → AI agent → human escalation → checkout → post-purchase thread — under real (not simulated) conditions before the campaign window opens. The dress rehearsal surfaces the gaps the load test didn't find: AI agent edge cases, escalation handover failures, in-thread recovery timing issues. Better to find them at day 70 than at 11pm on September 9th.
Daily marketing intelligence reads locked. The marketing intelligence layer that was built in month one now gets configured for live campaign reads — daily insight on what's working by channel, by audience, by creative, by chat conversion rate. The marketing head and the agency team have a shared view of campaign performance in real time, not in the weekly review meeting. Adjustment decisions happen on a 24-hour cycle, not a seven-day cycle.
In-thread cart abandonment recovery built and tested. When a buyer hesitates inside a WhatsApp or platform chat conversation — asks a question, gets an answer, and then goes quiet — the recovery message fires inside the same chat thread within minutes. Not an email the next morning. Not a retargeting ad two days later. An in-thread nudge, in the same conversation, while the buying intent is still alive. Attributed back through integrated D2C, this recovers a meaningful share of conversations that would otherwise have walked away.
By day 85, everything is live in a pre-warm state. The last five days before 9.9 are for monitoring, not building. That's the goal of the 90-day plan.
The marketing intelligence, conversation commerce, and integrated D2C work isn't what marketing agencies are built to do. The 90 days only fits if you have a parallel tech partner running the AI layer alongside the agency. Decide in week one — not week six.
The 30/60/90 plan above assumes a capable team for the AI layer exists from day one. In most APAC D2C marketing organisations, it doesn't. The marketing agency is structured for creative, media, and campaign management. The in-house team is structured for brand, strategy, and agency relationships. Neither is structured to build marketing intelligence infrastructure, wire conversation commerce AI agents, or build and test integrated D2C plumbing.
The 90-day plan collapses if the team question isn't answered by week one. Month one's three infrastructure tracks — marketing intelligence, conversation commerce, integrated D2C — all require a team that can build them in parallel, not sequentially. That team is the parallel tech partner, running alongside the agency, owned by the brand.
This is where RedDot Solutions fits. We are that parallel team. We build the marketing intelligence layer on the brand's first-party data — including customer intelligence — the conversation commerce infrastructure in WhatsApp and platform chat, and the integrated D2C plumbing that makes it all attributable. We sit next to the marketing agency, we don't replace it. The agency runs the campaign. We run the AI layer underneath it.
The decision to bring in a parallel tech partner is not a month two decision. It's a week one decision. Everything else in the 90-day plan — the brief timing, the creative spec, the attribution model, the pre-warm strategy — follows from it. Make it early. The 90 days fits if you do.
Three phases. Each builds on the last. The campaign goes live in phase three on infrastructure that already exists — not on assumptions.
| Phase | What You Build | What You Decide |
|---|---|---|
| Days 1–30 | Marketing intelligence on last year's data. Conversation commerce in WhatsApp + platform chat. Integrated D2C plumbing. | Who owns the AI layer. Parallel team in by day 14. |
| Days 31–60 | Agency brief built off marketing intelligence findings. Creative pipeline specced for variant volume. Ad-to-chat flow wired. Integration load-tested. | Hero asset vs variant volume split. Creative AI partner if agency can't ship variants. |
| Days 61–90 | Platform pre-warm. Conversation commerce dress rehearsal. Daily marketing intelligence reads. In-thread recovery loop. | Daily read cadence and escalation thresholds during campaign window. |
Where RedDot fits in this picture
RedDot Solutions is the parallel tech team for the AI layer of the 90-day plan. Marketing intelligence on first-party data. Conversation commerce inside WhatsApp and platform chat. Integrated D2C across Shopee, Lazada, and brand site. We sit next to your marketing agency, owned by you. We complete the team — we don't replace it.
Over the next few weeks I'll be publishing a few more pieces on this — what AI creative pilots actually look like in practice, what marketing intelligence means once it's set up properly, and a longer piece on conversation commerce specifically (because it deserves its own treatment, not a paragraph).
If you're a brand running on Shopee, Lazada, or D2C in APAC and any of this is hitting somewhere — these pieces are written for you.
Follow me here for the next pieces. Or read the full version on reddot.solutions.
If you have 90 days until 9.9 and you want to talk through the plan — specifically the parallel team question — we'd be happy to map it out together.
RedDot Solutions is the tech agency for APAC brands — marketing intelligence, conversation commerce, integrated D2C.
— Balan K
Founder, RedDot Solutions
Balan K is the founder of RedDot Solutions, an APAC technology partner working with D2C and consumer brands on AI-native marketing infrastructure. Based between Singapore and Chennai.
reddot.solutions