Malaysia Festive Ads: When to Launch (2028)

Fixed, movable, or sighted: the only split that matters

Most Malaysian marketing calendars are chronological lists. That ordering is useless for planning, because it hides the one thing that actually sets your lead time: how far in advance the date is knowable.

Sort the year into three groups instead. The groups never change, even though the dates inside them do, which is why this is the part worth memorising.

Fixed forever. Merdeka Day is always 31 August. Malaysia Day is always 16 September. Christmas Day is always 25 December. The X.X marketplace dates (9.9, 10.10, 11.11, 12.12) are numeric by definition. Black Friday belongs here too, but by rule rather than by number: it is the day after US Thanksgiving, which is the fourth Thursday of November. Every occasion in this group can be briefed, shot, designed, reviewed and scheduled a year out with literally zero date risk. Merdeka is always 31 August, so the brief lands in June and nothing that happens between now and then can move it.

Movable but known. Chinese New Year moves against the Gregorian calendar but is fixed by the lunar calendar far in advance. Deepavali is the same kind of movable, set by the Hindu calendar and published years ahead. You cannot recite these from memory, but you can look them up in July and trust what you find. The date is a lookup, not a forecast. The planning cost is one diary check, not a contingency.

Moon-sighted. The Islamic dates: Awal Ramadan, Hari Raya Aidilfitri, Hari Raya Aidiladha. These are declared when a sighting committee confirms the new month, so any Gregorian date printed ahead of time is an estimate. The tell is that reputable trackers print different days for the same occasion. That is not sloppiness, it is an honest reflection of a date that does not exist yet.

That split is the whole planning system. Group one gets a long, calm runway and a locked schedule. Group two gets the same runway plus a diary check. Group three gets a window, a flexible schedule, and creative that does not hard-code a date.

Everything below applies the system to the 2028 run. The dates will age. The groups will not.

Why a warm-up phase is not optional

Here is the argument for launching early, built entirely from things you can check.

Criteo's commerce data, reported by eCommerce News Asia, found that Malaysian retail sales rose 26% in the final two weeks of Ramadan 2025, peaking at a 52% uplift on 23 March. Across Southeast Asia, retail sales rose 13% year on year during Ramadan 2025, with shoppers beginning product discovery earlier in the month while conversions still clustered toward Eid. So the money arrives at the end. That part everyone knows, and it is why so many Malaysian brands treat the last ten days as the campaign.

The number that should change your plan is the other one. Across the six Southeast Asian markets Criteo measured (Indonesia, Malaysia, Singapore, the Philippines, Vietnam and Thailand), purchases completed in those final two weeks came after an average gap of 19 days between a shopper's first product visit and the purchase. That is a regional average rather than a Malaysia-only figure, but the +26% Malaysian pattern above sits inside it: the browsing that produces a Raya sale starts roughly three weeks before the sale happens. If your first impression lands in the last week, you are not in that consideration set. You are competing for the fraction of demand that decides on impulse, at the exact moment the auction is most expensive.

Now add the mechanical delays, which stack on top of the shopper's timeline rather than overlapping with it.

Meta's ad review starts automatically before ads begin running and is typically completed within 24 hours, although it may take longer in some cases. That is Meta's own published wording, and "may take longer" is doing real work in it. You cannot plan a peak-day switch-on against a best case.

Then the learning phase. An ad set stays in the Facebook learning phase until it reaches roughly 50 optimization events in a seven-day period, and any change to creative counts as a significant edit that restarts learning (WordStream, 2024). A campaign launched on sale day spends the sale learning instead of converting. If your product converts a handful of times a day, fifty events is not a formality, it is a week or more of spend.

Stack the three: a 19-day regional average from first visit to purchase, an ad set that needs roughly fifty events to stabilise, and a review queue that is usually a day but not guaranteed. The warm-up phase has to be live weeks before the peak, not days.

That conclusion is ours, not a citation. No publisher states a weeks-ahead figure for Malaysian festive ads, so treat the recommendations below as planning judgement reasoned from the facts above, and adjust them to your own conversion volume.

Lead time by group, then by occasion

Start with the group-level rule, because it survives every calendar change.

Now the 2028 run, with dates verified against the trackers cited at the foot of this page. Weeks-ahead means when the warm-up goes live, not when the brief starts.

The pattern in that column is worth naming, because it is the second evergreen rule on this page. Runway tracks consideration length, not the size of the sale. Deepavali and Raya need long runways because people are choosing gifts and outfits and comparing them over weeks. 12.12 needs a short one because the same shopper already met your brand at 11.11 and Black Friday. A big sale with a short decision cycle does not need a long build. A modest sale with a four-week decision cycle does.

Warm-up phasing, in every case, means the same shape: the early weeks carry product and brand content that earns cheap reach and seeds the consideration set, and the peak window carries the offer. The two are different creative, not the same ad at a higher budget.

That shape has a budget consequence worth stating in ringgit. If the Raya peak is carrying RM10,000, the warm-up weeks should already have spent something like RM3,000 to RM4,000 seeding the consideration set before the offer ever runs. Treat that as an allocation shape rather than a benchmark: it is the split we would plan to, not a published Malaysian figure, and there is no verified local festive CPM to calibrate it against. The point is that the warm-up is a real line in the budget, not the leftovers. A plan that puts the entire festive budget behind peak week is not a phased campaign, it is a one-week campaign with a long brief. Our companion pieces on seasonal versus evergreen Facebook ads and Facebook ad CPM in the holiday season cover why the split matters for the auction.

Fixed forever: the campaigns you can brief a year out

Merdeka Day is 31 August, every year, without exception. Malaysia Day is 16 September. Neither has ever moved and neither will.

That makes them the easiest campaigns in the Malaysian year to run well, and the ones most often left to the last fortnight. There is no excuse for a rushed Merdeka ad. You could have shot it in January. Since the date carries no risk, put the effort into the thing that actually differentiates Merdeka creative, which is tone rather than timing, and read our Merdeka and Malaysia Day ad playbook for the angles that hold up.

The only variable is the weekday, and it is worth ten minutes of planning. In 2028, Merdeka falls on Thursday 31 August, so the sentiment build runs through the working week and spills into the long weekend behind it. Malaysia Day lands on Saturday 16 September, a weekend peak with a natural family-time audience rather than a mid-week commute. Same fixed dates, different phasing, and you only know which by checking the weekday each year.

The marketplace dates need no forecasting at all. 9.9, 10.10, 11.11 and 12.12 are numbers. What they need is an understanding of how the marketplaces themselves phase them, because your ads run in the same feed.

Malaysian marketplaces do not launch on the day. In 2025, TikTok Shop ran an on-ground 11.11 Mega Sale Carnival from 7 to 9 November at Pavilion Bukit Jalil, and Lazada ran a 28-hour marathon sale starting 10 November with hourly RM11 deals and a KOL-led livestream kicking off at 8PM that night, while TikTok Shop reported over 200% year-on-year growth in brands joining its Mall program (Marketing-Interactive, 2025). The peak is a finish line, not a starting gun. If the platforms are already selling on the 7th, an ad that switches on at midnight on the 11th arrives after the demand has been captured. The mega sale ad playbook covers the offer construction for those dates.

Movable but known: Chinese New Year and Deepavali

These two behave identically from a planning standpoint, and the discipline is one calendar check.

Chinese New Year 2028 falls on Wednesday 26 January, with a second day on Thursday 27 January (PublicHolidays.com.my). Deepavali 2028 falls on Tuesday 17 October, a public holiday nationally except in Sarawak. Neither is a forecast. Both were knowable years ago and neither will move.

What does move, year to year, is the gap between them and everything else. That is where the planning work actually is.

Take Deepavali. In 2026 it landed on Sunday 8 November, three days before 11.11, which meant the Deepavali peak, the marketplace warm-up and the 11.11 warm-up all crowded into one week with no production time between them. In 2028 it falls on 17 October, roughly three clear weeks before 11.11, which is a comfortable handover instead of a scramble. Same occasion, same group, completely different production calendar. The lesson is not "Deepavali collides with 11.11", it is "check the gap every year and let it set your sprint". The Deepavali ad playbook has the occasion-level angles once the timing is settled.

Chinese New Year has the same property in reverse. In 2028 it collides with something else entirely, which is the next section.

The collision that moves every year

Awal Ramadan 2028 is estimated at Friday 28 January (PublicHolidays.com.my), pending moon sighting. Chinese New Year 2028 is Wednesday 26 January and Thursday 27 January. Those are two days apart. The second day of CNY runs straight into the eve of Ramadan.

Read that again if you sell to both audiences, because it changes your production calendar rather than just your media plan. In a normal year you run CNY, take a breath, brief Ramadan, build it, launch it. In 2028 the breath does not exist. Whatever your Ramadan-into-Raya campaign is going to be, it has to be briefed, produced, approved and loaded into the ad account before CNY week starts, because the moment CNY ends you are already late. Assume January 2028 is your production month for both campaigns, not one. The CNY ad playbook covers the creative half.

Two honest caveats. The Ramadan date is an estimate, because it officially begins only when a moon-sighting committee declares the new month. And Malaysia's public holiday calendar for a future year is not final until the Cabinet publishes it: the government's own portal points readers to the Cabinet's public holiday list at kabinet.gov.my, which is what you should re-confirm against before you commit budget. Neither caveat rescues you. Even if the sighting pushes Ramadan a day, CNY and Ramadan are still inside the same week.

Here is the part that generalises. This collision is not a fixture. The Islamic calendar runs about eleven days shorter than the Gregorian one, so Ramadan and Raya walk backwards through the seasons year after year, colliding with a different occasion each time and then moving on. The permanent rule is not "watch out for CNY and Ramadan". It is: every year, before you plan anything, print your two or three biggest occasions on one page and measure the gaps between them. The gap, not the date, is what decides whether you have one production sprint or two.

Moon-sighted: build Raya against a window, not a date

This is where the discipline is genuinely different, and the evidence for it is unusually clean.

Hari Raya Aidilfitri 2028 is estimated at 27-29 February, and the trackers do not agree. PublicHolidays.com.my lists Sunday 27, Monday 28 and Tuesday 29 February. Public-Holidays.com.my lists Sunday 27 and Monday 28 February. TrueCalendar lists 27-28 February and marks the Islamic dates tentative and subject to moon sighting. Reputable sources, a day apart, for the biggest commercial event in the Malaysian calendar.

That disagreement is the honest reflection of a date that does not exist yet. Ramadan officially begins when a moon-sighting committee declares the new moon and new month, and Raya follows from it. Hari Raya Aidiladha 2028 is estimated at Friday 5 May on the same basis. Every one of those is an estimate, and will remain one until the announcement.

So plan the uncertainty in rather than disclaiming it.

Build the campaign against the window, and make sure every asset works on any date inside it. Set the ad schedule to the earliest date, because being live a day early costs you one day of budget while being live a day late costs you the peak. Never hard-bake a date into the creative image itself: a burnt-in "28 February" is a re-render, a re-upload and a fresh review queue if the sighting moves, and Meta keeps ads subject to review and re-review at all times. Put any date that must appear into the ad copy, where it is a text edit, or better, into a countdown framed in days rather than dates. And re-confirm the Cabinet's public holiday list before the media budget is committed.

None of that is a compromise. A window-built campaign is not a worse campaign, it is the same campaign with the date risk engineered out of the artwork. The occasion-level arc lives in the Ramadan ad playbook and the Raya ad playbook, with the creative-compliance side in the halal ad creative guide. The year-by-year date reference sits in the Malaysian marketing calendar.

Black Friday as a bridge, and the year-end run-out

Black Friday is fixed, but by rule rather than by number. It is the day after US Thanksgiving, which is the fourth Thursday of November, putting it on Friday 24 November in 2028 (PublicHolidays.com). Derive it from the rule when you plan a future year rather than assuming an ordinal, and it is knowable as far ahead as you like.

Look at where that sits: after 11.11, before 12.12. That position is the whole strategy. Black Friday in Malaysia rarely justifies its own ground-up creative build, because your audience met your brand at 11.11 thirteen days earlier and will meet it again at 12.12 eighteen days later. Treat it as a bridge. Keep the 11.11 audiences warm, refresh the offer, run a shorter runway, and save the production budget for the two peaks either side. Our Black Friday ads guide for Malaysia covers the offer construction.

On cost, be careful what you import. Gupta Media's tracker found the week including Black Friday (ISO Week 48, 25 November to 1 December) was the most expensive week of 2024 at an average Meta CPM of USD 13.42, and recorded a Cyber Monday CPM of USD 17.70. It put Meta's Q4 premium against Q1 to Q3 rates at 13% in 2022 and 17% in 2023. That is US and global data, not a Malaysian benchmark, and no verified Malaysian festive CPM exists in any neutral source. Read Gupta's figures for shape (Q4 costs more, the Black Friday week is the spike) and measure your own ringgit CPM against your own off-peak baseline rather than against a foreign average.

Then Christmas Day, always 25 December, closes the year. In 2028 it lands on a Monday, which stretches the run-out across a long weekend and pulls the gifting deadline forward: the last confident delivery date, not the 25th, is what your ads should be counting down to. The year-end and Christmas ads guide has the creative side.

How to research past timing in the Meta Ad Library

You do not have to guess when your category switches on. Last year's answer is sitting in public.

The method is narrow and it works. Search the brand in the Meta Ad Library, open Filters, and use "Impressions by Date" to select a single day or a date range going back up to seven years (Shopify, 2026). Then ask one question per brand: on what date did last year's Raya or 11.11 creative first appear? Step the range backwards a few days at a time until the festive ads stop showing, and the boundary you hit is the switch-on date. Do that for five or six brands in your category and you have a real distribution of when your competitors actually started, rather than when they say they started.

Be clear about what this cannot do. Meta describes the Ad Library as a searchable database of all active ads, but publishes spend, reach and funding entity data only for ads about social issues, elections or politics. There is no spend figure for a commercial festive campaign. Anyone telling you what a Malaysian brand spent on 11.11 from the Ad Library is inventing it. Timing, and creative, are what the library can honestly teach you.

The second limitation is retention. Meta stores political ads in the Ad Library for seven years, and EU-targeted ads are archived for one year after their last impression. Commercial ads carry no equivalent published retention guarantee. A competitor's Raya ad that is paused today may simply not be there next January when you go looking. So screenshot the timing evidence when you find it, with the date range visible in the frame. Keeping your own dated screenshots, or working from a searchable archive of Malaysian ads such as AdPlay.ai, beats trusting that the query still returns anything a year from now. The discipline of capturing the evidence when you see it matters more than where you keep it.

What you should expect to find, if 2025 is any guide, is that the phasing is real. Marketplaces were on the ground from 7 November and running marathon sales from the evening of 10 November for an 11.11 peak. The brands doing this well were visible well before the date on the poster.

The mistakes that cost a peak

Four failure modes, each with a fix that costs nothing except doing it earlier.

Launching on the day. The campaign spends the highest-competition day of the quarter in review and then in learning. Fix: upload days early, schedule the switch-on, and let the ad set exit learning on cheap warm-up traffic.

Swapping creative on peak morning. Meta keeps ads subject to review and re-review at all times, states that ads edited to bring them into compliance are treated as new ads, and any creative change restarts learning (WordStream, 2024). Fix: build the peak creative as new ads in a scheduled ad set rather than editing the warm-up ads in place.

Hard-baking a sighted date into an image. Works fine until the sighting moves, then it is a re-render and a fresh review at the worst possible moment. Fix: dates live in copy, or count down in days.

Carrying last year's collisions forward. Deepavali sat three days before 11.11 in 2026 and nearly a month before it in 2028. CNY and Ramadan are two days apart in 2028 and will not be by 2030. Fix: measure the gaps between your biggest occasions at the start of every planning year, and let the gaps, not your memory, set the production sprints.

Bringing it together

The Malaysian festive calendar is not hard to know. It is hard to act on early enough, because the fixed dates feel too far away to work on and the sighted dates feel too uncertain to work on. Both instincts are wrong, and in opposite directions.

Merdeka on 31 August, Malaysia Day on 16 September, Christmas on 25 December, Black Friday behind Thanksgiving and every X.X date are certainties. Brief them a year out and there is no downside. Chinese New Year and Deepavali are lookups, not forecasts: check them once, then plan them like fixed dates. Only Ramadan and the two Rayas carry live sighting risk, and the correct response to that is a window, date-free imagery and an earliest-date schedule, not a delayed start.

Meanwhile the shopper is already moving. A 19-day regional average from first visit to purchase, in a market with 23.0 million Facebook users and 16.1 million Instagram users against 35.4 million internet users (DataReportal, 2026), means the consideration that produces your peak-week revenue is happening while you are still approving the artwork. Get the warm-up live, keep the offer for the peak, and check the Ad Library for what your category did last year before you decide you know better.

Then re-confirm your dates against the Cabinet's public holiday list at kabinet.gov.my, which the government's own portal points to. A future year's calendar is never final, and the whole point of a long runway is that you have time to react when it changes.