↖︎ Vishal Singh

The Political Ad Ledger · Chapter 2

The Anatomy of a ₹383-Crore Ad Election

Seven years of Google political ads in India — 397,522 of them — reveal a machine that wakes only for elections, spends four of every five rupees on video, and is dominated by a single party running still images at industrial scale.

In the calendar year of India's 2024 general election, verified political advertisers spent ₹383 crore on Google's ad platform — roughly what they had spent in the previous five years combined. This is not a story about steady growth. It is a story about a switch. India's digital political advertising is almost entirely a function of the election timetable: near-silent between contests, then a wall of spending in the weeks a vote is held. Every rupee figure here comes from Google's own transparency data, which reports India spending only in rupees and only as ranges; we take the midpoint of each range.

Across the whole window, the pattern is stark. Annual spend climbs from a negligible ₹2.2 crore in 2019 to ₹383 crore in 2024, then falls back — before rising again in 2026 as a fresh round of state elections arrives. The advertising does not track the news cycle, the economy, or the season. It tracks polling days.

1  The calendar, drawn in ad money

Plot weekly spend across seven years and the elections draw themselves. Every local maximum is a campaign: the 2024 Lok Sabha peak (the week of May 5, 2024 alone saw ₹33.5 crore from 121 advertisers), the November 2024 Maharashtra and Jharkhand assembly polls, and — at the right edge — the live 2026 state-assembly cycle already spiking. Between elections, the line all but flatlines.

Weekly political-ad spend on Google in India

Sum of per-ad spend midpoints, INR crore per week · vertical guides mark major elections

Data table (top weeks)
Figure 1. Weekly spend, Feb 2019 – Jul 2026. Spend is Google's reported range midpoint summed across all ads active that week; it is an estimate, not an audited figure. The 2024 Lok Sabha campaign is the tallest structure in the series by a wide margin. Hover for the week's spend and active-advertiser count.

The height of these spikes has grown election over election, but their shape is constant: a fast ramp in the six weeks before a vote, a peak in the final fortnight, and an almost immediate collapse afterward. Political ad spending in India is bursty by construction — the ad accounts exist to win specific contests, not to maintain a presence.

2  Three-quarters of the ads, one-fifth of the money

What are advertisers buying? Overwhelmingly, in count, static images: 298,091 of the 397,522 creatives are image ads. But images are cheap. Measured by spend, the picture inverts completely — video absorbs about four of every five rupees, on barely a quarter of the ads. Text ads are a rounding error. The strategic split is clear: images for cheap, tileable reach; video for the money shots.

The format inversion: share of ads vs share of spend

Every India creative by format · left bar = share of the 397,522 ads · right bar = share of spend

Data table
Figure 2. Image ads are three-quarters of the volume but a fifth of the spend; video is the reverse. Video is also targeted differently — 60% of video ads narrow by age or gender, versus 4% of image ads (hover a segment for its targeting rates). The money and the demographic aiming travel together.

3  Most ads are whispers

The impression counts (which Google also reports only as ranges) show how bottom-heavy the distribution is. Roughly 40% of all India political ads never cleared a thousand impressions. A small number of national video spots reach into the tens of millions, but the median creative is a micro-targeted whisper aimed at a single constituency — consistent with the fact that 89% of ads geo-narrow to a state or smaller.

How far each ad actually reached

Ads grouped by reported impression band · count of creatives per band

Data table (raw buckets)
Figure 3. Impression bands collapsed into decades of reach. The sub-1,000 band alone holds ~40% of all creatives; the ≥10-million band is a thin slice of expensive national video. Impressions are Google's reported ranges, binned by their midpoints.

4  One party, industrial scale

Nine parties account for most party spending, but one sits entirely apart. The Bharatiya Janata Party ran 274,587 ad creatives — thirty-eight times the Indian National Congress's 7,270 — yet only 9% of them are video. It is a saturation strategy: enormous volumes of cheap, geo-targeted images. Congress runs a fraction of the ads but leans video (40%). The regional parties cluster tightly in time around their own state elections — Biju Janata Dal's entire run is the 2024 Odisha window; Bharat Rashtra Samithi's is a single month of the 2023 Telangana vote.

Volume versus video: how each party advertises

One dot per party · x = number of ad creatives (log) · y = share that are video · area = spend

Data table
Figure 4. The BJP is the lone outlier at the bottom-right: an order of magnitude more ads than anyone, almost all of them images. Bubble area is spend (midpoint estimate). Hover any dot for the party's full profile.

5  Where the money lands

Spending is not distributed like population. Tamil Nadu leads every state at ₹141 crore — well ahead of Maharashtra and West Bengal — a consequence of ferociously contested, ad-heavy Dravidian-party politics. The geography of Google political ad money is really a map of where digital campaigns are fought hardest, not where the most people live.

Total political-ad spend served, by state

From Google's geo_spend table (all advertisers combined) · INR crore

Data table (all states)
Figure 5. The 16 highest-spend states/UTs. This is the only geographic breakdown Google publishes for India — total spend served into each region across all advertisers; there is no per-advertiser state table for India (that exists only for the United States).

A state total says who was reached, not who did the reaching. Line the biggest advertisers up by how many states each one targets and a second geography appears: the BJP blankets the country — its ads touch 69 states and union territories — while regional parties are rifles, not shotguns. The Biju Janata Dal advertises in exactly one state (Odisha); the YSRCP in two. National ambition and regional focus are written into the reach of a single advertiser.

How many states each big advertiser reaches

Distinct states / UTs an advertiser's ads target · the twelve largest advertisers by spend

Figure 6. National parties and the Union government blanket dozens of states; regional parties concentrate on home turf (BJD: Odisha only). Consultancies sit in between — I-PAC's ~24 reflect its multiple clients. Hover for the count.

6  The money is concentrated

For all its 877 advertisers, this is a market ruled by a handful. The top 10 advertisers account for 71% of all spend; the top 50, for 92%. The BJP alone is nearly 30%. And on the reach side the format logic sharpens: images buy far more eyeballs per rupee than video — about 36 impressions per rupee against video's 11 — which is exactly why the volume game is played in images and the money is spent on video.

The twelve advertisers that hold most of the money

Spend of the largest advertisers · ₹ cr · coloured by type · the top 10 = 71% of all spend

Figure 7. The BJP (≈30% of everything), then the government, the Congress, and the consultancies. Past this handful, spend falls away fast — the top 50 of 877 advertisers hold 92%. Hover for type.

Put together, the anatomy is coherent. A political-ad system that switches on for elections and off between them; that buys cheap images for breadth and expensive video for impact; whose reach is mostly local and small, with rare national blasts; dominated by one party operating at a scale no one else attempts; and concentrated in the states where the fights are closest. None of it required reading a single advertisement — only counting them.

Data & methods

  • Source. Google's bigquery-public-data.google_political_ads dataset (the public mirror of the Ads Transparency Center), filtered to ads verified for India. 397,522 creatives from 877 advertisers, first shown Feb 2019, extracted 14 July 2026.
  • Spend is a bucket midpoint. Google reports India spend only in INR and only as ranges (e.g. ₹2.5M–₹3M). Every spend figure here sums the per-ad range midpoint; treat totals as estimates with a band around them, not audited amounts. For the ~0.06% of ads in an open-topped range, we use the lower bound.
  • Impressions are ranges too, binned by midpoint. Weekly and state totals use Google's own point estimates (the advertiser_weekly_spend and geo_spend tables); the per-format and per-party figures use the creative-level midpoints, so the two do not sum identically.
  • "Political" is Google's classification. The dataset is advertisers Google verified as political; its net also catches some commercial brands and all levels of government (explored in companion pieces). Party rows here are the nine largest by spend.
  • Honest limits. The window is Google only — not Meta, not television, not print. It captures digital display advertising on Google/YouTube, a slice of total campaign spending. The 2026 figures are partial (through mid-July).