The Diwali App Marketing Playbook

The festive season is not a demand opportunity you capture. It is an auction you either enter prepared or lose money in.

Most festive advice optimises for winning installs. The teams that come out ahead optimise for what those installs are worth in December. Those are different objectives and they produce different plans.

Diwali 2026 falls on Sunday 8 November. This is the whole picture, with pointers to the detail on each part.

What is the Indian festive season from a growth team's view?

Diwali app marketing is the practice of planning, running and measuring mobile app campaigns across India's festive advertising window, which runs from the late September sale events through Diwali in November and is characterised by compressed budgets, elevated auction costs and lower-intent install volume.

The window has three waves, not one:

Whether you should participate depends on a distinction most teams skip:

• Real demand lift: commerce, quick commerce, payments, gifting, travel, gaming top-ups. Festive genuinely changes what your users want to do.

• Cost lift only: most content, social, productivity and edtech apps. You pay the higher auction prices without receiving the demand.

If you are in the second group, spending your annual peak budget in your least efficient month is a choice rather than an obligation. Seasonal businesses need an MMP more than year-round ones because the measurement stakes are higher for concentrated windows, regardless of which group you are in.

The cost reality

Across roughly 4.7 million installs at around 50 India-based apps running paid campaigns on Linkrunner between May and July 2026, the median blended cost per install was just under Rs20, with the middle half between roughly Rs8 and just under Rs40.

That is the non-festive baseline. Festive moves it, and it moves for four reasons that compound: auction density as every large advertiser bids at once, budget compression as annual money is spent in six weeks, faster creative fatigue under high frequency, and attention shifting into shopping apps rather than yours.

The more important point is that cost per install is a poor guide to whether a festive campaign worked. Install volume rises during festive while install quality falls, so CPI can look flat or even improve while your real cost climbs.

In the same India sample, apps in the bottom quartile by activation rate paid close to three times their CPI for each activated user. Apps in the top quartile paid roughly 1.2 times. Same metric, entirely different meaning.

Full cost breakdown, including how to set a walk-away number and structure reserve-and-release budgeting, is in the dedicated post on festive UA costs. The general levers are in our guide to optimising CPI while maintaining quality.

Planning: the 90-day sequence

Working backwards from 8 November:

• Days 90 to 60 (August to early September). Agree the budget envelope and the walk-away number. Make the vertical call. Capture your baseline. Verify and freeze measurement. Brief creative, do not produce it yet.

• Days 60 to 30 (September to early October). Test creative while inventory is still affordable. Treat Big Billion Days as a live rehearsal. Set scaling rules in writing.

• Days 30 to 7 (October). Scale only what passed testing. Release budget in tranches. Hold 15 to 20 per cent in reserve.

• Final week (2 to 10 November). Daily checks only. No structural changes.

• Days 1 to 30 after. Retention watch, reconciliation after 14 days, and a one-page handover.

The single concept worth carrying from this sequence is the measurement freeze deadline: a date, around mid September, after which nothing about how you measure changes. Everything about how much you spend stays flexible. The full week-by-week calendar is in the dedicated planning post.

Measurement: set it up before you spend

This is the part most festive guides append as a closing paragraph, and it is the part that decides whether any of the rest is knowable.

Festive traffic breaks setups that work all year, because volume spikes expose race conditions that low volume hides, because a broken week is a large share of a six-week campaign, and because fraud follows the value of the inventory.

One specific failure mode deserves naming here. Big Billion Days runs roughly ten days. Your default click-through attribution window is probably seven. A user who taps your ad on day one and converts on day nine is unattributed or credited elsewhere, so the campaign that created the demand looks worse than it performed, and may get cut mid-window on that basis. There is no correct setting, but there is a difference between deciding this in September and discovering it in December.

The pre-festive checklist covers nine items: SDK freeze, per-network postback verification, Meta Install Referrer configuration on Android, deep link QA against new festive surfaces, revenue event testing, reattribution handling, fraud thresholds, volume alerting and baseline capture. Each is detailed in the dedicated attribution setup post. If you only run one thing beforehand, run the weekly attribution audit in early September.

Creative and campaign structure

Two principles carry most of the weight.

Test before the auction reprices. Every concept you plan to scale should be validated in September, when the same test costs a fraction of what it costs in Diwali week and teaches you the same thing.

Expect faster fatigue and plan rotation, not replacement. Frequency climbs during festive, so creative burns out in days rather than weeks. The answer is a validated set to rotate within, not a new concept introduced at peak CPM. The creative fatigue guide covers the detection thresholds.

On structure, the automation-versus-control question matters more during festive because the learning phase is expensive when CPMs are high. Our comparison of Meta Advantage+ Shopping and manual campaigns sets out when each is appropriate. Make sure your festive landing surfaces route correctly, particularly if you are driving web-to-app traffic from sale pages.

What festive installs are actually worth

The November number and the January number tell different stories, and only one of them is true.

Two things distort the November view. First, a share of your "new" installs are returning users being re-credited to a new campaign, which inflates install volume and lowers reported CPI without any real efficiency gain. Second, festive-acquired new users, reached under broad targeting into a discount-driven moment, tend to be the weakest cohort of your year and their retention decline shows up in December.

Blend the two populations and you get a curve that describes nobody. Split them by cohort at the point of arrival and you get two honest curves and a real answer about whether to repeat this next year. The dedicated post on reattribution versus new installs covers the split, and attribution data can drive the retention response for each group separately.

The checklist

By mid August: budget envelope agreed, walk-away cost per activated user set, vertical decision made, baseline captured, creative briefed.

By mid September: measurement verified and frozen, creative tested, scaling rules written, attribution window decision made and recorded.

October: scale validated creative only, release budget in tranches, hold 15 to 20 per cent reserve.

2 to 10 November: daily checks on installs, activation rate and the network-to-MMP gap. No structural changes.

After 24 November: reconcile, watch festive cohort retention separately from baseline, write the one-page handover.

If you want to see what your own festive cohorts cost per activated user rather than per install, request a demo and we will walk through the cuts on your data.

FAQ

When does app marketing for Diwali need to start?

Around 90 days out, which for 2026 means early to mid August. Budget decisions, vertical calls and measurement freezes all need lead time, and creative has to be tested in September while inventory is still affordable.

Do festive installs retain as well as normal installs?

Usually worse. Festive installs are acquired under broad targeting into a discount-driven moment, which produces the weakest cohort of most apps' years. The decline appears in December and January rather than in the festive dashboard, which is why cohorts need tagging when they arrive.

Which app categories benefit most from festive campaigns?

Categories with genuine festive demand: commerce, quick commerce, payments, gifting, travel and gaming top-ups. Most content, social, productivity and edtech apps see cost lift without demand lift and often do better buying the shoulder weeks.

How much should I increase my budget for Diwali?

Rather than picking a multiplier, set a maximum tolerable cost per activated user and release budget in tranches against evidence from each spend wave, holding 15 to 20 per cent unallocated so you can scale what works in the final week.

What is the most common festive app marketing mistake?

Judging the campaign on install volume and cost per install. Both improve during festive for reasons unrelated to performance, which is why teams repeat expensive campaigns. Cost per activated user, tracked by festive week, is the number that answers the question.