Meta & Google Performance Marketing: Interview Playbook for Structures, Creative Testing and Marginal ROAS
A shopper does not wake up and say, “I will enter a funnel now.” She scrolls Instagram, sees a frame style she likes, forgets it, later searches “blue light glasses under 2000,” compares options on Google, and finally buys after a remarketing nudge. The best performance marketers do not treat Meta and Google as two ad dashboards - they treat them as one demand system.
- Meta creates and harvests demand visually through feeds, Reels, Stories and catalog ads; Google captures intent through Search, Shopping, YouTube and Performance Max.
- A strong account structure separates prospecting, retargeting, brand search, non-brand search and shopping/catalog so budgets and learnings do not get mixed.
- Creative testing is not “try many ads.” It is a controlled system: define hypothesis, isolate variable, spend enough, read full-funnel metrics, then scale winners.
- ROAS = revenue from ads / ad spend; it is useful, but dangerous if treated as truth without margins, attribution and incrementality.
- Marginal ROAS tells you what the next rupee of spend returns. Scale only while marginal ROAS stays above your economic hurdle rate.
- The common trap is optimizing for platform-reported ROAS and accidentally overfunding retargeting or brand search, which often claims credit for demand created elsewhere.
Big Picture
The Meta and Google playbook has three linked decisions: where the campaign sits in the customer journey, what creative or keyword hypothesis it tests, and whether the next unit of spend is still profitable. If you remember one mental model, remember this left-to-right flow.
The Playbook: Structure, Test, Then Scale by Marginal ROAS
1. Start with the role of each platform
Meta is strongest when the product benefits from visual discovery, social proof, creator-style content or impulse consideration. It can create demand before the customer has a search query.
Google is strongest when the customer already has intent: a keyword, a comparison query, a product search, a local need or a YouTube research moment. It captures demand that is already forming.
2. Build an account structure that matches the buying journey
A clean structure prevents the classic mistake: letting high-intent audiences and low-intent audiences sit in the same budget pool. You want the algorithm to learn, but not to hide where profit is coming from.
During large fashion sale periods, a marketplace like Myntra typically needs both discovery and capture. Meta-style short video and carousel creatives can stimulate browsing for categories such as sneakers or ethnic wear, while Google Search and Shopping capture shoppers actively searching brand, sale and category terms. The strategic point: the win comes chiefly from matching channel role to customer intent, supported by merchandising depth, app retargeting and sale-period urgency.
3. Test creative like a system, not like a lottery
On Meta, creative is often the largest lever because targeting has become broader and more automated. On Google, creative still matters through ad copy, assets, landing pages and product feed quality, but keyword intent and feed relevance carry more weight.
A good creative test isolates one hypothesis: the hook, the offer, the proof, the format or the audience pain point. If five things change at once, you may get a winner but not a learning.
4. Read the right metrics before scaling
Performance dashboards can seduce you with precision. A 4.2x ROAS looks scientific, but it may hide thin margins, repeat buyers, retargeting credit or delayed returns. Use a compact scorecard.
5. Use marginal ROAS to decide whether to scale
Average ROAS tells you what the whole campaign returned. Marginal ROAS tells you what the next slice of budget returned. Scaling decisions should depend more on the second one.
Worked example: when to stop scaling
Suppose a D2C backpack brand has 55% gross margin and 10% variable fulfilment, payment and return costs. Contribution before ads is therefore 45% of revenue.
Break-even ROAS = 1 / contribution margin = 1 / 0.45 = 2.22x.
If an extra ₹1,00,000 of ad spend generates ₹2,60,000 incremental revenue, marginal ROAS is 2.6x, so scaling still makes sense. If the next ₹1,00,000 generates only ₹1,90,000 incremental revenue, marginal ROAS falls to 1.9x, below break-even. The right decision is to stop, refresh creative, improve conversion rate or shift budget.
Definitions You Must Say Cleanly
Performance marketing: Paid digital marketing optimized against measurable actions such as clicks, leads, purchases or revenue.
ROAS: Revenue attributed to advertising divided by advertising spend.
Marginal ROAS: Incremental revenue generated by the next unit of advertising spend divided by that incremental spend.
Incrementality: The additional outcome caused by marketing that would not have happened without the marketing activity.
Creative testing: A structured experiment comparing ad variants to learn which message, format or offer improves business outcomes.
Lenskart: Balancing Demand Creation and Demand Capture
Lenskart shows how an Indian omnichannel brand can use Meta-style visual discovery and Google-style intent capture without confusing the two roles.
Situation: Eyewear is a considered purchase. Customers care about style, lens type, price, prescription accuracy and trust. In India, many buyers also like offline reassurance: trying frames, checking fit and getting eye tests. That means one channel rarely closes the whole journey.
The move: A performance playbook for a brand like Lenskart works by giving each channel a different job. Meta-style creatives can show frame styles, face shapes, offers and lifestyle use cases. Google Search and Shopping can capture high-intent queries such as eyeglasses, sunglasses, computer glasses or brand-specific searches. App, CRM and retargeting then bring back product viewers, cart abandoners and repeat buyers.
Outcome and lesson: The primary driver is journey fit: discovery for style inspiration, search for active intent, and omnichannel trust for conversion. Supporting drivers include a wide product catalog, first-party customer data, store network, app experience and retargeting discipline. The strategic lesson is clear: do not ask Meta and Google to perform the same job; make them pass demand to each other and judge the whole system on marginal profitability.
How AI Changes The Meta & Google Performance Playbook
AI does not remove the need for marketing judgment. It moves the marketer’s work from manual lever-pulling to better inputs, cleaner experiments and sharper economic decisions.
Automated bidding can optimize toward the conversion signal you give it, even if that signal is commercially weak. If you feed it low-quality leads, repeat-buyer-heavy conversions or inflated attribution windows, AI will scale the wrong outcome faster.
Interview Relevance
“A D2C brand says its Meta ROAS is 3.5x and Google ROAS is 6x. The founder wants to move most of the budget to Google. How would you evaluate this decision?”
Say this line in the interview: “I would not move budget from Meta to Google just because Google has higher reported ROAS; I would first separate brand demand from incremental demand and compare marginal ROAS.”
Common Mistake
The costly mistake is treating platform-reported ROAS as the final truth. It can over-credit brand search, retargeting and repeat buyers while under-crediting demand creation. One-line fix: use ROAS for diagnosis, but use incrementality and marginal ROAS for budget decisions.
What to Revise Next
Next, revise Conversion Rate Optimization (CRO) Fundamentals because better landing pages raise the value of every paid click. After that, move to Retention, CRM & Lifecycle Marketing so you can connect acquisition cost with repeat purchase, LTV and profitable growth.