Performance marketing is simple to describe and hard to do: put money in, get more money out, know exactly why. We run Meta, Google, LinkedIn and marketplace ads with one obsession, profitable scale, and one rule, the numbers get reported whether they flatter us or not.
What we run
- Meta ads: prospecting, retargeting and creative testing systems
- Google ads: search, shopping, Performance Max and YouTube
- LinkedIn ads for B2B pipelines
- Marketplace advertising on Amazon, Flipkart, Myntra and Nykaa
- Landing page CRO so clicks stop leaking at the last step
- Tracking architecture: pixel plus server-side, clean conversion data
Why accounts plateau, and how we fix it
Most stuck accounts share three problems: the algorithm is learning from dirty data, creative fatigues faster than it gets replaced, and the landing experience was never built for the traffic. We rebuild all three as one system.
Clean signals first
Server-side tracking and conversion audits so every rupee of learning budget teaches the algorithm something true.
Creative as a system
10 to 15 ad variants per cycle with clear kill and scale rules. Creative is the biggest lever in paid, so we treat it like one.
Full-funnel accountability
Prospecting, consideration and closing each get their own creative, offer and measurement. Blended ROAS hides problems; funnel-stage reporting exposes them.
For a women’s fashion D2C brand, this system cut CPA by 30% while scaling spend. Read the case study. A lifestyle brand grew sales 400% on the same playbook. That one too.
Creative is the only lever you still control
Meta and Google have automated almost everything that used to be a media buyer’s job. Targeting is largely algorithmic. Bidding is automated. Placement is decided for you. Broad targeting with a strong creative now routinely beats a carefully constructed audience stack, and the platforms are explicit that this is the direction of travel.
What remains in your hands is what the ad actually says and shows, and how many versions of it you can put into the system. That makes performance marketing a production problem far more than an optimisation problem, which is uncomfortable for agencies built around dashboard management.
The practical consequence is unglamorous but decisive: a brand producing forty assets a month has a structural advantage over one producing four, regardless of who is better at media buying. More concepts tested means faster learning, and creative fatigue stops being fatal because there is always something new entering rotation. This is the single biggest reason our media and production teams sit in the same building.
The measurement problem nobody wants to discuss
Platform-reported numbers are not neutral. Meta and Google each attribute conversions to themselves, using their own windows and their own logic. Add the two dashboards together and you will frequently find they claim more revenue than your bank account received.
Signal loss made this worse. iOS restrictions, cookie deprecation and consent requirements mean a meaningful share of conversions are modelled rather than observed. The dashboards do not stop reporting when the data stops arriving; they estimate, and they estimate in their own favour.
So we report blended numbers alongside platform numbers, and we say which is which. Blended acquisition cost, total marketing spend divided by total new customers, is the figure that ties to your accounts. It is less flattering than in-platform ROAS, which is precisely why it is useful. If a month looks good in the dashboard and flat in the bank, we will say so before you have to ask.
Where performance accounts actually break
- Scaling before the economics work. If contribution margin does not cover acquisition cost, more spend accelerates the loss. This is the most common and most expensive failure we are called in to fix.
- Too few creatives. Running one asset until cost per acquisition visibly climbs means you have already overpaid for weeks. Fatigue shows up in the numbers long after it shows up in the feed.
- Optimising to the wrong event. Add-to-cart and lead-form submissions are easy for the algorithm to find and often uncorrelated with revenue. You get exactly what you ask for, which is rarely what you wanted.
- Audience fragmentation. Fifteen small ad sets starve the algorithm of the conversion volume it needs to learn. Consolidation feels like losing control; it is usually the fix.
- Judging brand spend on last-click. Awareness work will always look bad in a ROAS column. That does not mean it is not working, and it does not mean every rupee of it is working either.
- Broken tracking nobody checked. A duplicated pixel, a missing server-side event or a consent banner blocking fires will quietly corrupt months of learning. We audit this before we touch a budget.
How we structure an account
Consolidated campaigns with enough conversion volume to exit the learning phase, not a sprawl of micro-targeted ad sets. A clear separation between prospecting and retargeting budgets so one cannot quietly cannibalise the other. Naming conventions that survive a team change. Creative testing on a fixed cadence with pre-agreed kill and scale thresholds, so decisions are made by rule rather than by whoever looked at the dashboard last.
Server-side tracking alongside the browser pixel, with conversion events that map to revenue rather than to activity. Offline and post-purchase data fed back where the business allows it, because the algorithm can only optimise toward what it can see.
The first ninety days
Weeks one and two: audit and instrumentation. Account structure, tracking integrity, conversion event mapping, creative library, landing page performance and unit economics. You get the findings whether or not you continue with us.
Weeks three to six: rebuild and learn. Restructured campaigns, clean signals, a first full creative cycle in market, and a weekly report from the first week, including the weeks that do not flatter anyone.
Weeks seven to twelve: scale what survives. Winning concepts get iterated into variants and formats; losers get cut without ceremony. Budget follows evidence, and the blended number is the one we hold ourselves to.
What we need from you
Honest unit economics, including contribution margin after shipping, returns and discounts. Access to the data that tells us whether a lead became revenue. A decision-maker who can approve creative quickly, because slow approvals cost more than most media inefficiencies. And a tolerance for being told when paid media is not your problem, which is more often than the industry admits.
What budget do we need to start?
Enough for the algorithm to learn and for tests to reach significance. For most Indian D2C brands that means a monthly media budget from around 1.5 to 2 lakh; below that we will tell you honestly whether paid is your best first move.
How fast will we see results?
Expect 2 to 4 weeks of structured learning, then optimization begins to bite. We report weekly from day one, including the ugly weeks.
Do you also make the ad creatives?
Yes, in-house. Statics, motion, UGC-style and product films. Creative and media sit at one table here, which is exactly why the system works.
Meta or Google: which is better for us?
Wrong question, and we say that kindly. The right split depends on your category, margins and intent landscape. The audit tells us; the data then re-tells us monthly.
Do you charge a percentage of ad spend?
No. A percentage of spend rewards us for spending more, which is the opposite of the incentive you want. We work on a flat monthly fee based on scope, so the advice to slow down costs us the same as the advice to scale.
Can you work alongside our in-house team?
Often the best setup. Common splits are in-house owning brand and organic while we own paid media and performance creative, or in-house running day-to-day while we handle creative production and strategic review. We will tell you which parts you should keep in-house rather than sell you all of it.
What if our margins do not support paid media?
Then we say so in the audit and tell you what would have to change first, usually pricing, average order value or retention. We would rather lose a retainer than scale an account that loses money faster with every rupee added.
How is marketplace advertising different?
On Amazon, Flipkart, Myntra and Nykaa you are buying intent that already exists, so the levers are listing quality, keyword coverage, ratings and price competitiveness as much as bids. We treat marketplace as its own discipline rather than a Google campaign in a different interface.