A DTC ecommerce brand spending in the mid-five-figure range on Google Ads each month watched its ROAS decline for three consecutive quarters despite its in-house marketing manager working longer hours and making more frequent adjustments. The problem was not effort. It was infrastructure. After transitioning to a done-with-you Google Ads management model with groas, the brand restructured its campaigns, rebuilt its shopping feed, and recovered purchase revenue within six months. This is the story of how that happened, why the original setup was always going to plateau, and what other ecommerce brands at similar scale can learn from it.
A Google Ads ROAS recovery for ecommerce is rarely about finding one broken lever. It is about diagnosing compounding structural issues that manual management cannot outrun. This case study walks through the diagnosis, the execution, and the results.
The Situation: A Growing Shopify Brand With Shrinking Returns
This brand sells a consumer product line through Shopify, generating the majority of its revenue through Google Ads. Monthly ad spend hovered around $40K. The business was growing, the product-market fit was strong, and repeat purchase rates were healthy.
The in-house marketing manager ran everything: Search campaigns, Shopping, Performance Max, remarketing. He was experienced enough to build campaigns, write ads, adjust bids, and troubleshoot Merchant Center issues. He was doing all the right things at the tactical level.
But the numbers told a different story. ROAS had dropped from over 5x to under 3x across three quarters. CPCs were climbing. Performance Max was consuming more budget but producing fewer conversions at a worse cost per acquisition. The marketing manager was spending more time in the account, not less, and the returns were still heading in the wrong direction.
The founder started exploring agencies. Most wanted onboarding fees north of $5K and six-month commitments. The founder was not ready to hand over the keys entirely, and the marketing manager did not want to be sidelined. What they wanted was better infrastructure and a strategic partner, not a replacement.
The Diagnosis: Three Problems Compounding Each Other
When the groas team onboarded the account, the first step was not making changes. It was a full diagnostic. The weekly reporting and strategy call cadence that comes with the DWY model meant the in-house manager stayed in the loop at every step. What the diagnostic revealed was not a single broken campaign. It was three structural problems feeding off each other.
Over-Segmented Campaign Structure Creating Internal Competition
The account had 11 active campaigns, many of which targeted overlapping audiences and keywords. Multiple Shopping campaigns competed against each other for the same product categories. Search campaigns had ad groups that cannibalized each other’s traffic. The result: the brand was bidding against itself in auction after auction, driving up its own CPCs while splitting conversion data so thin that Google’s algorithms could not optimize effectively.
This is a common pattern with in-house management. The instinct is to create more granular control by adding campaigns. But in Google’s auction system, more campaigns does not mean more control. It means more fragmentation. Understanding why this ceiling forms is the first step toward fixing it.
tROAS Targets Set Too High, Starving Campaigns Of Volume
The marketing manager had set aggressive target ROAS values, thinking this would force Google to find only the most profitable conversions. In practice, it did the opposite. With tROAS set at 600%, campaigns had almost no room to explore. They served fewer impressions, accumulated fewer conversions, and the algorithm’s learning phase never completed. The campaigns were stuck in a low-volume loop where the data was too sparse to improve performance.
This is one of the most counterintuitive dynamics in Google Ads: high ROAS targets often shrink revenue instead of protecting it. The fix is not to remove targets entirely but to calibrate them against the volume the account actually needs to generate enough signal.
A Shopping Feed That Was Working Against The Account
The product feed had not been touched in months. Titles were generic, missing high-intent search terms that shoppers actually use. Custom labels were absent, meaning there was no way to segment products by margin, sell-through rate, or promotional status. Product type taxonomy was shallow and inconsistent, which limited how well Google could match products to relevant queries.
A poor feed does not just reduce Shopping performance. It degrades Performance Max, which pulls from the same Merchant Center data. Every campaign that used product feeds was operating on bad inputs. Feed quality is one of the most underrated ROAS levers in ecommerce, and in this account, it was actively dragging everything down.
The Decision: Why Done-With-You, Not Full Handoff
The founder considered the DFY model but ultimately chose DWY for a specific reason: the in-house marketing manager was capable and motivated, and the founder wanted to stay informed on strategy without being in the weeds.
The groas DWY model fit this situation precisely. The proprietary engine, trained on over $500 billion in profitable ad spend, runs underneath doing the heavy lifting on execution. A senior strategist works alongside the in-house team, providing a weekly report on exactly what was done plus a strategy call every other week. The in-house team stays in the driver’s seat. groas provides the infrastructure, the data advantage, and the strategic layer that one person working alone cannot replicate.
Onboarding was $0. No contract lock-in. Month-to-month, cancel anytime. The brand started within days, not weeks.
The Execution: What Changed And Why
The first 30 days were about structural repair, not tactical tinkering. Here is what happened and the reasoning behind each change.
Campaign Consolidation: From 11 Campaigns To 4
The groas strategist and the in-house manager worked together to collapse 11 campaigns into 4 with clean budget allocation. The goal was to stop internal competition, consolidate conversion data, and give Google’s bidding algorithms enough volume per campaign to optimize effectively.
The four remaining campaigns were built around clear roles: one broad Shopping campaign segmented by custom labels, one Performance Max campaign restructured by product category with layered audience signals, one branded Search campaign, and one non-brand Search campaign focused on high-intent category terms.
This consolidation alone reduced wasted spend from internal auction overlap within the first two weeks.
Feed Overhaul: Titles, Custom Labels, And Merchant Center Diagnostics
The groas engine and strategist rebuilt the product feed from the ground up. Product titles were rewritten to include specific product attributes, brand name, and search terms that matched how shoppers actually query. Custom labels were created to segment products by margin tier, seasonality, and promotional status. Product type taxonomy was rebuilt to give Google clear category signals.
Merchant Center diagnostics were run to flag disapprovals, data quality warnings, and competitive benchmarks on price and availability. Several products that had been suppressed due to policy issues were fixed and reactivated.
The impact of feed quality on Shopping impression share was visible within weeks. When Google understands what you sell, it shows your products to more relevant shoppers at better CPCs.
Bidding Strategy: Relaxing tROAS To Recover Volume
The tROAS target was reduced from 600% to 350% across the core campaigns. This was not a permanent change. It was a deliberate move to allow campaigns to accumulate conversion data, exit the low-volume loop, and let the algorithm learn which audiences and placements actually convert.
The groas engine monitored performance continuously, adjusting bids and budget allocation 24/7 rather than relying on the in-house manager to check in during business hours. As conversion volume recovered, the strategist and the in-house team worked together on the biweekly calls to recalculate the right tROAS threshold based on actual margin data, not aspirational targets.
PMax Restructuring With Audience Signal Layering
The original Performance Max setup had a single asset group targeting all products with no audience signals. The restructured version used distinct asset groups organized by product category, each with audience signals built from first-party customer data, competitor audiences, and in-market segments relevant to the specific product line.
This gave PMax the structural clarity it needs to perform. Performance Max is powerful, but only when the inputs are precise. Without clean feeds, proper asset group segmentation, and layered signals, it defaults to spending budget on low-quality placements.
The Results: Six Months After Transition
Within the first 60 days, ROAS stabilized and began climbing. By month four, purchase revenue had recovered to its previous peak. By month six, the account was generating meaningfully more revenue at a higher ROAS than its best quarter before the decline began.
Average CPC dropped after campaign consolidation eliminated internal competition. Shopping impression share increased as the feed improvements gave Google better data to match products against relevant queries. Performance Max shifted from a budget drain to the highest-volume, most efficient campaign in the account.
The in-house marketing manager’s role changed. Instead of spending hours each day making manual bid adjustments and troubleshooting campaigns, he focused on creative strategy, product launches, and coordination with the groas strategist on upcoming promotions and seasonal shifts. The groas engine handled the continuous optimization. The strategist brought the expertise and the broader pattern recognition from working across hundreds of accounts. The in-house manager brought the business context and product knowledge that no outside team can fully replicate.
That combination, the engine running 24/7, a senior strategist providing advisory and oversight, and the in-house team contributing business context, is what the DWY model is designed to produce.
What This Means For Ecommerce Brands At Similar Scale
The core lesson from this brand’s experience is not that they were doing Google Ads wrong. They were doing it the way most in-house teams do: manually, with good intentions, and without the infrastructure to scale.
The difference between managing Google Ads and optimizing the engine that runs it shows up when accounts hit a certain level of complexity. Below $10K per month in spend, manual management can work. At $40K per month with multiple product lines and channels, the number of variables exceeds what one person can optimize in a business day. The gap is not about skill. It is about throughput and data advantage.
When DWY Is The Right Fit
If your team has someone who knows Google Ads and you want to keep them in the driver’s seat, DWY gives you the infrastructure upgrade without losing control. You get the groas engine running underneath, a senior strategist working alongside your team, and a cadence of reporting and strategy calls that keeps everyone aligned. This is where this brand started, and it is where most ecommerce brands with competent in-house marketers should start.
When DFY Makes More Sense
If the founder is the one running Google Ads, or the in-house person is stretched across multiple channels and cannot commit to acting on strategic recommendations, DFY is the better path. groas owns everything end to end, including landing pages and offers. Nothing to log into or manage. There is no shame in admitting that Google Ads deserves a dedicated owner, and groas is built to be that owner.
Customers often start on DWY and upgrade to DFY as they scale or as the founder gets pulled into other priorities. The strategist flags the upgrade when the timing makes sense.
What This Brand Would Do Differently From Day One
Looking back, the founder said the biggest mistake was assuming that having someone capable in the account was the same as having the right infrastructure under it. The marketing manager was never the problem. The ceiling was structural: fragmented campaigns, starved bidding algorithms, and a feed that was working against every campaign it touched. Those problems do not get solved by working harder. They get solved by changing the foundation.
For any DTC ecommerce brand spending at a level where ROAS trajectory matters to the business, the question is not whether your person is good enough. It is whether the infrastructure underneath them is good enough. If it is not, you will hit the same plateau this brand hit, and adding hours will not fix it.
groas puts a senior strategist on top of a proprietary engine trained on hundreds of billions in ad spend so that execution does not stop when a human runs out of hours. The gap shows up in the numbers inside the first few weeks. Month-to-month, no onboarding fees, cancel anytime. If you have an in-house team and want the engine plus a strategist, get started with DWY. If you want Google Ads fully handled, apply for DFY and the team will figure out the right plan on the call.
