I’ve been saying something in rooms for the last few months that gets a reaction every time: digital is going to become as expensive as every other advertising medium. The discount is closing. Even in India. And creativity will need at least double what you’re budgeting for it right now.
People push back on that. Then they go and check their own numbers, and come back quieter.
The promise digital was originally sold on was arbitrage — cheaper reach than print, better measurement than television, no minimum spend to get started. Two of those three are already gone. The third is going.
But I want to do something slightly uncomfortable in this post: argue against my own headline. I’ve now spent enough time with the 2026 data to know that the cost argument, while directionally right, is the weaker of the two reasons you should be building brand presence today. And if you take your case to the board on the weaker argument, you’ll lose it.
Let me show you both.
First, the part I got right
The cost inflation is real, and it’s measurable.
Meta’s average CPM crossed $14 in 2026, up roughly 20% from $11.82 the year before. Average cost per acquisition across industries now sits near $38 — up around 38% from 2025. That’s not a seasonal blip. That’s the auction repricing itself.
Here’s the mechanism most people miss. Advertisers are pulling money out of Google Search because AI Overviews have gutted the click-through economics there. Seer Interactive tracked 3,119 queries across 42 organisations and found organic CTR fell 61% when an AI Overview appeared — from 1.76% to 0.61%. Paid CTR fell 68%. So the budget migrates to Meta, where CPMs look cheaper. Everyone does this at once. Meta’s auction gets crowded. CPMs rise.
You’re not paying more because you got worse at this. You’re paying more because the room filled up.
India isn’t exempt. Digital ad spend here grew 19% in 2025 to ₹71,621 crore, per Dentsu’s Digital Advertising Report, and is tracking toward ₹98,034 crore by 2027. YouTube CPMs in India have been climbing 15–20% year on year since 2024. Our CPMs are still low in absolute terms — around $2.60 versus $23 in the US — but the direction is unmistakable, and the gap is closing faster than most Indian founders have priced in. The cheapness that made digital feel like a free option for Indian brands was a market condition, not a law.
And the creative point? I actually understated it.
Meta’s own data puts creativity at roughly 56% of a campaign’s conversion outcome — more than targeting, bidding and placement combined. That’s because Advantage+ and broad targeting have collapsed the old targeting advantage entirely. Creative is the targeting now. The algorithm needs variation to find the pockets of buyers inside an audience it has already chosen for you.
Which means the volume requirement has changed shape completely. Current benchmarks put a brand spending $10,000 a month at 15–20 active creatives, with the creative lifecycle compressed to 14–21 days. At $50,000 a month across Meta, TikTok and Google, you need 15–25 variants running at once just to hold CAC flat. Motion’s 2026 benchmark report — 550,000+ ads, $1.3 billion in spend — found that only about 6% of ads capture the majority of that spend, and that for every ten creatives tested, one to three become real winners.
Most brands ship two to four creatives a month. The requirement is four to ten times that.
So no, doubling your budget isn’t enough. Volume is the entire game now.
Now, the part I’d argue differently
Here’s where I take my own case apart.
The two-year timeline is the shakiest thing I said, at least for India. CPMs here, compounding at 15–20% a year, don’t close a nine-fold gap with Tier 1 markets inside twenty-four months. The arithmetic doesn’t get there. Parity is coming, but for Indian brands the honest window is longer than the one I’ve been quoting — and I’d rather correct that here than have someone correct it for me.
There’s also a countervailing force I was ignoring. AI has made production radically cheaper. A workflow that took a mid-size brand a month to produce 8–15 assets can now produce that in an afternoon. One good operator with the right stack does what needed five people in 2023.
Required volume is up sharply. Unit cost of production is falling sharply. The net direction is genuinely ambiguous, and anyone who tells you otherwise is selling something.
There’s a sharper risk hiding inside that, and it’s one my own industry is exposed to: if execution cost keeps collapsing while required volume keeps rising, what breaks isn’t the price point — it’s activity-based pricing itself. “₹X per month per activity” assumes the activity stays the billable unit. That assumption is under more threat than the number attached to it.
But the deeper problem with the cost argument is strategic, not arithmetic.
Money can always buy media. At 2x, at 3x, at 5x. A founder in 2028 with a bigger budget than yours can walk in and outbid you tomorrow morning. So “it will be expensive later” isn’t actually an argument for starting today. It’s an argument for having more revenue later.
If cost is your only reason to start now, a well-funded competitor beats you with a cheque.
The argument that actually forces “today”
Money cannot buy retroactive presence.
That’s the whole thing. That’s the sentence I’d build the strategy on.
There are assets in digital that are cumulative and time-locked, and no amount of budget in 2028 buys back your absence in 2026. Branded search volume history. Citation density across the sources language models actually pull from. Review corpus depth. Entity consistency across the web — the same name, the same description, the same claims, in enough places, for long enough, that machines treat you as a known thing rather than a candidate.
You cannot compress that. You can only start it.
And the data on what that presence is worth is now embarrassingly strong.
Harvard Business Review’s 2026 research found companies in the top 10% for brand awareness carry 3.1x lower customer acquisition cost than the market average. Les Binet’s IPA analysis found branded search volume is the single strongest predictor of future revenue growth — a 10% increase in branded search predicts 5–8% revenue growth. Brand awareness shortens the B2C sales cycle by around 24%.
Then there’s the number that should genuinely change how you allocate budget this quarter. Authoritas’s 2026 SERP study found that Google’s AI Overviews mention well-known brands 2.3 times more often than unknown ones.
Read that again in the context of everything above.
The machines that now sit between your customer and their decision favour brands that already exist in the corpus. Zero-click rates on AI search products run between 60% and 93%. Nearly two-thirds of Google searches end without a click. The click is disappearing as a unit of value, and what’s replacing it is being mentioned — being the name the model reaches for.
You can pay to be seen. You cannot pay to be remembered by a system that was trained before you showed up.
Binet and Field’s IPA Databank — 1,400+ campaigns over 30 years — established that brand effects need six months minimum to start compounding. WARC’s Multiplier Effect data found roughly a 90% ROI uplift moving from performance-only to brand-plus-performance, and a 40% ROI decline going the other way.
Six months to start. Two years to matter. That’s the actual clock, and it has nothing to do with CPMs.
Also, Check – Should You Hire a Fractional CMO or an Agency?
What this means for how you spend
Three things I’d hold to.
One: brand is the only line item whose CAC contribution deflates while paid inflates. Every other input gets more expensive per unit. Presence gets cheaper per unit acquired, the longer you’ve had it. That asymmetry is the entire reason the reinvestment loop works.
Two: measure the loop, or you’re running on faith. This is where most founders lose the argument internally. Brand spend dies because the performance team looks at it through last-click attribution, sees no purchase events, and cuts it. Six months later, CAC is up and there’s no equity to lean on.
So instrument it properly. The metrics that prove brand is compounding are branded search volume, direct traffic, share of voice, citation frequency in AI answers, and blended CAC or MER. If your paid CPMs are climbing while your blended CAC holds flat or falls, the loop is real, and you should feed it more. If blended CAC climbs in lockstep with CPM, you don’t have a brand — you have a media buy with a logo on it.
That distinction is worth more than any budget-split framework.
Three: stop treating creative as a project. It’s a production line now, with a required throughput. If you’re shipping four assets a month against a 15-variant requirement, you don’t have a creative problem. You have a capacity problem wearing a creative problem’s clothes.
So, was I wrong?
About the conclusion, no. About the reasoning, partly — and I’d rather say that publicly than have you build a two-year plan on the softer half of my argument.
Start now, but not because it gets expensive. Costs move in both directions, and anyone putting a precise date on parity two years out — including me — is guessing.
Start now because presence is the one input that can’t be bought in arrears. Because the systems deciding who gets recommended are weighting existence over spend. Because the compounding curve has a minimum runway, and it starts the day you begin, not the day you decide it’s important.
You can outbid someone for attention. You cannot outbid them for the two years they’ve already spent existing.
Where to go from here
If you’re reading this and quietly recognising that your paid costs are climbing while your branded search line is flat, that’s the diagnostic. That gap is the whole story, and it’s fixable — just not quickly.
I work with a small number of founders and CMOs each quarter on exactly this problem: building the brand and AI-visibility layer that makes performance spend cheaper over time, and putting the measurement in place to prove it’s working before the CFO asks.
If that’s the conversation you need to be having, write to me at hello@mansirana.com with two things — your blended CAC trend over the last four quarters, and your branded search trend over the same period. I’ll tell you honestly whether you have a brand problem or a media problem. They need very different amounts of money.
The brands that will be affordable to grow in 2028 are being built right now, by people who decided not to wait for the case to become obvious.
Don’t be the one running the numbers in 2028 and realising the cheapest thing you could have bought was time.