BeatXP: Modeling a 3.2x ROAS on a Simulated Rs30L Campaign

For my MyCaptain capstone I built a full paid-media plan for BeatXP's Flare Pro Smartwatch: a simulated Rs30L budget across Meta and Google, modeled to a 3.2x ROAS and an 18% lower cost per lead. It was a coursework simulation, not a live client account, but the planning, structure, and math are exactly how I approach a real launch. Here is how it came together.

The brief

BeatXP was launching the Flare Pro Smartwatch into a saturated wearables category where prior campaigns hovered around a 1.4x ROAS. The goal I set: model a plan that could realistically clear 3x ROAS on a Rs30L budget across Meta and Google in an eight-week window, without leaning on discounting to buy the numbers.

Funnel architecture

I split the budget across three stages so each rupee had a job:

• Awareness: broad interest targeting plus 1% lookalikes on Meta, and Display plus YouTube on Google, to seed a retargeting pool.
• Consideration: retargeting engagers and site visitors with proof-led creative (specs, comparisons, reviews).
• Conversion: custom audiences and Advantage+ Shopping on Meta, Search plus Performance Max on Google, where the intent and the budget weight were highest.

Search and PMax carried the conversion load; Display and YouTube existed to fill the top of the funnel, not to be judged on last-click ROAS.

Creative testing

Creative is where most performance budgets quietly leak, so I built a 4x3 testing matrix: four hook angles (price, feature, social proof, aspiration) across three formats (single image, carousel, video reel). Each variant had rotation triggers tied to CTR and CPM, so fatigued creative got cut before it dragged the account average down.

Budget and bid discipline

The plan modeled dynamic budget reallocation every 72 hours based on ROAS and CPA thresholds: scale what clears the bar, trim what does not. On Google, Target CPA bidding paired with negative-keyword hygiene and audience-signal enrichment kept spend on qualified intent. The 18% modeled CPL reduction came almost entirely from this discipline, killing underperforming ad sets quickly rather than hoping they would recover.

The modeled results

To be clear on scope: these are modeled projections from a capstone media plan, delivered as a full Notion build, not live-account results. What is real is the method, and it is the same method I would run on a live budget.

What actually transfers

The spreadsheet is disposable. The system is not: give every stage of the funnel a distinct job, test creative on a fixed matrix instead of on vibes, and reallocate budget on a schedule so winners scale and losers die fast. That loop is what turns a saturated-category launch from 1.4x into 3x-plus.

Planning a launch or trying to lift a paid account that has plateaued? Here is how I work.

Further reading

Deeper dives on the pieces above, from my Marketing Academy:

• Meta Ads
• Google Shopping & Performance Max
• Creative testing
• Bidding strategies