In advertising the user isn't the customer. The audience is the inventory. Revenue = attention (users × time) × monetisation (ad load × price per unit of attention). Costs split two ways. One part manufactures the attention: content for Sun TV, an index of the web for Google, creators for YouTube. The other is traffic acquisition, wherever reach is rented. The signature economics follow from that. One more impression costs close to nothing, which gives you spectacular incremental margins once the attention machine is built. And the whole thing is pro-cyclical. Ad budgets are the first line CFOs cut, so ad revenue amplifies the macro cycle, a link Phase 7 will make.
Advertising business model example: Alphabet
Alphabet (US). At an illustrative ~2024 scale (pull the 10-K), revenue is ~$350bn. Search is ~$198bn of it, YouTube ads ~$36bn and Network ~$30bn, so ads together come to ~$265bn. Cloud adds ~$43bn and subscriptions/devices make up the rest. Two leaks define the waterfall. The first is TAC (traffic acquisition cost), ~$55bn paid to Apple, Samsung, browsers, and network partners for default placement and distribution. About 21 cents of every advertising dollar goes out so the attention arrives pre-aggregated. The Google-Apple default arrangement is worth roughly $20bn/yr to Apple per the DOJ trial record. It's Google's TAC leak and Apple's licensing archetype at the same time. One contract, two anatomies. The second leak is the modern one. Capex was ~$52bn in 2024 and has been guided far higher since. AI compute has turned an asset-light auction house into an industrial capex machine. Operating margin is ~32%, but from here on watch FCF conversion as well as EBIT.
Advertisement
How does the advertising model work in India?
India. Digital is now over half of Indian ad spend, in a ~₹70,000+ cr market where Google and Meta capture the large majority of digital (check the current GroupM/dentsu figure). That leaves the listed Indian pure-plays mostly adjacent. Sun TV is a broadcaster, with ad revenue of ~₹1,200 to 1,300 cr inside ~₹4,000 cr of revenue and PAT margins >40%. It's a lesson in what happens to beautiful ad economics when the attention pool itself, linear TV, structurally shrinks. Affle is ad-tech that charges CPCU, cost per converted user, which makes it an advertising model priced like usage. When you analyse any ad business, work through three checks in order. First, is the attention pool growing or draining? Second, is time × ad load near saturation? Last, check whether pricing (CPM/CPC) is rising because more bidders have joined the auction, or because the format mix is shifting (video vs banner).
What should you check in the filings?
What to check in filings: user and engagement metrics against ad revenue growth, since price-led growth with flat engagement has a ceiling. The TAC ratio and who it's paid to, because that's a dependence. Ad revenue by format and property. Look at the 2009 and 2020 quarters for the cyclicality history. And since 2023, capex as % of revenue, plus the depreciation shadow it casts into future COGS.