Forwrd

HYDRAID • Case Study • Google Ads

From low 5-figure conversion value to 7 figures. Revenue grew nearly twice as fast as spend.

An account that had failed to grow under multiple previous attempts was rebuilt from first principles, unlocked from its own constraints and transformed into a full-funnel growth engine across the DACH market.

The numbers

Comparing March 2024 – March 2026 against the equivalent prior period.

0% Conversion value growth — from low 5 figures to 7 figures

0x Revenue grew faster than spend, improving efficiency as the account scaled

0x Growth in generic search monthly revenue

About Hydraid

Hydraid is a German hydration brand built around a scientifically formulated carbohydrate-electrolyte powder developed according to WHO guidelines for oral rehydration solutions. Designed to hydrate faster than plain water through the body's sodium-glucose co-transporter mechanism, Hydraid targets athletes, runners, cyclists, and active people who need reliable, low-sugar, evidence-based hydration. Made in Germany, vegan, and free from artificial colours and flavourings, Hydraid competes in a category dominated by major sports drink brands and increasingly by Amazon marketplace sellers.

Core account truth

The problem was structural, not tactical.

The constraints themselves prevented the data density needed to improve. The path forward was blocked by the same structural decisions that had created the problem in the first place.

The story

Multiple attempts. The same outcome.

Hydraid had tried multiple times to make paid search work before coming to us in February 2024. Each attempt had followed the same pattern: campaigns launched with tight ROAS targets, constrained bidding and a structure that prioritised efficiency over learning. The result was always the same too, low volume, minimal growth and an account that never built the momentum it needed to scale.

We took a different view entirely. Rather than optimising within the same constraints that had already failed, we reset the strategy from the ground up, removing the restrictions that were preventing the algorithm from doing its job, rebuilding the account architecture around volume and learning and introducing full-funnel channel coverage for the first time.

The low-learning loop

Restrictive ROAS targets

Low volume

Poor learning signals

Inefficient bidding

The account cannot learn

The challenge

Previous campaigns had been built around restrictive ROAS targets that, while theoretically sound, created a self-reinforcing trap: low ROAS targets produced low volume, low volume produced poor learning signals, poor learning signals produced inefficient bidding and inefficient bidding produced low ROAS.

The algorithm couldn't break out of this loop because the constraints themselves prevented the data density needed to improve.

The Amazon challenge

All of Hydraid's products are available on Amazon, which means that Google Ads campaigns compete not just against other DTC brands but against Hydraid's own marketplace listings.

In a category where price comparison is common and Amazon's trusted ecosystem captures a significant share of purchase intent, winning on Google requires a strategy that reaches buyers earlier in their journey, before they've defaulted to Amazon as their destination.

Reach buyers earlier

Hydraid DTCBuild preference before the final purchase decision

Earlier funnel
entry point

AmazonThe default destination for established purchase intent

The strategic reset

We started from first principles. Rather than inheriting the existing structure and optimising within it, we rebuilt the account architecture entirely, shifting from an efficiency-first framework to a scale-first framework and from constrained algorithmic operation to learning-led growth.

Testing, filtering and learning animation

Removing the low-learning loop

The single most impactful change was removing the restrictive ROAS caps that had prevented the algorithm from generating meaningful data. By allowing the system to operate at higher volume, accepting a wider range of conversion outcomes in the early phase, we gave the algorithm what it needed to learn: signal density, audience diversity and enough conversion events to build reliable bidding models.

Once the learning phase was complete, efficiency followed naturally from a better-informed system rather than being imposed artificially from the start.

Rebuilding search around genuine intent

Generic and category-led search campaigns were rebuilt from scratch around the actual language of Hydraid's target audience: electrolytes, sports hydration, runner hydration, cyclist recovery, and category-level terms that captured demand at different stages of the purchase journey.

DSA campaigns were introduced to ensure no relevant search intent was left uncaptured. Each campaign was structured to avoid cannibalisation with Shopping and PMax — channels doing complementary but distinct jobs.

Search intent and account optimisation animation

Distinct channel roles

SearchGenuine category and problem-led intent

ShoppingProduct-led demand and purchase readiness

PMaxPrimary revenue engine without forced overlap

Demand GenEarlier reach and future conversion pipeline

Shopping and PMax as revenue engines

Shopping and Performance Max were positioned as the primary revenue-generating channels structured deliberately to prevent forced overlap between them and the search campaigns.

The goal was a clean account architecture where each channel captured a distinct slice of demand rather than competing for the same conversions. This avoided the bidding inflation that occurs when multiple campaign types chase the same user and ensured that budget was allocated to where it was genuinely incremental.

Demand Gen for incremental reach

To reach buyers earlier in their consideration journey and specifically to intercept users who might otherwise default to Amazon, we introduced Demand Gen campaigns split between prospecting and remarketing.

Prospecting targeted new audiences who hadn't yet encountered Hydraid. Remarketing re-engaged users who had visited but not converted. Together, these campaigns extended the account's reach beyond high-intent search and built a pipeline of future conversions that pure search activity alone could never generate.

Prospecting and remarketing split animation

The results

Comparing March 2024 – March 2026 against the equivalent prior period.

Metric	Before	After	Change
Impressions	Low baseline	Massively expanded	+1,287%
Clicks	Limited volume	High volume	+414%
Ad spend	Minimal	Scaled investment	+2,479%
Conversion value	Low 5 figures	7 figures	+4,196%
Revenue vs spend growth	—	Revenue grew ~2x faster than spend	ROAS improved at scale
Generic search (monthly)	Negligible	8x higher (most recent month)	+8x

The headline number, conversion value growing more than 40x while spend grew roughly 26x, tells the most important part of the story: the account didn't just get bigger, it got more efficient as it scaled. That combination, volume and efficiency improving together, is the hallmark of a structural rebuild rather than simply spending more money.

Generic search alone captures the trajectory clearly. From a negligible monthly output before the reset to 8x that figure in the most recent month, reflecting both the expanded keyword coverage and the stronger commercial intent capture that comes from a correctly structured, unconstrained search account.

Key learnings

ROAS constraints can be the enemy of ROAS improvement.

Restricting an algorithm to a tight efficiency target before it has enough data to hit that target sustainably is one of the most common errors in paid search management. The way to improve long-term ROAS is to allow short-term learning, not to constrain it. Once the system has built reliable bidding models from real data, efficiency follows.

Amazon competition requires an earlier funnel entry point.

For a brand whose products are available on Amazon, winning on Google isn't primarily about capturing bottom-of-funnel demand, Amazon often wins that battle by default. The strategic advantage lies in reaching buyers earlier, building brand familiarity and ensuring that when the purchase decision is made, Hydraid's own channel is the preferred destination. Demand Gen was the structural answer to this challenge.

Channel architecture matters as much as channel presence.

Running Search, Shopping, PMax, and Demand Gen simultaneously only creates compounding returns when the channels are structured to avoid cannibalisation. Clean separation of channel roles, each owning a distinct part of the funnel, is what makes a multi-channel strategy genuinely additive.

Scale and efficiency are not opposites.

When the starting point is a constrained, low-learning account, removing the constraints often improves both volume and efficiency simultaneously, because the algorithm was never operating at its potential to begin with.

The result

From low 5-figure conversion value to 7 figures. Revenue growing nearly twice as fast as spend. Generic search scaling 8x in monthly output.

An account that had failed to grow under multiple previous attempts, rebuilt from first principles, unlocked from its own constraints and transformed into a full-funnel growth engine across the DACH market.

The previous strategies weren't wrong to want efficiency. They were wrong to prioritise it before the conditions for efficiency existed. Sequence matters in paid search: build the learning base first, then let the efficiency emerge from it.

Ask me about a free audit

Meet Yann Durand

Co-founder of Forwrd. Google Ads, YouTube Ads and performance marketing specialist.

Hi, I'm Yann. I've spent the last 10+ years living and breathing Google Ads, YouTube and performance marketing — and honestly, I still love it.

I started out at a leading global agency in the UK after a first-class master's in digital marketing. Then I moved to the Netherlands to run Google Ads for Bestseller on brands like Jack & Jones, Vero Moda and Only, spending millions across 16 brands and 15 markets simultaneously.

The chapter that really changed things was Vienna. At waterdrop, I took non-brand Google Ads from €200k to €6M in revenue. That experience became the backbone of a system for D2C brands trying to scale with Search, Shopping and YouTube Ads.

After seeing the bad shape of most Google Ads accounts out there, I built Forwrd to go after the untapped potential most agencies leave behind. Where agencies stay on the surface, we go after the moves that actually shift revenue.