Select Page
Case Study: Hero

Case studies / eCommerce

Google Shopping & feed recovery

Two thousand products. None of them eligible to appear.

PowerPak Civil & Safety had twenty-five years of reputation in construction PPE and a Merchant Center feed that wouldn't approve. Shopping wasn't underperforming: it couldn't run. The bidding was never the problem.

Client PowerPak Civil & Safety
Sector Industrial PPE · B2B + B2C
Catalog 2,000+ SKUs
Engagement 6 months · Congers, NY
Case Study: Results
900% Return on ad spend Six months
$250K+ Attributed revenue In-period
$1.14 Average cost per click Across 23,000 clicks
4.6M Impressions earned From zero eligible
Case Study: About the Client
PowerPak Civil Safety
About the client

Two thousand products. None of them eligible to appear.

PowerPak Civil & Safety is a New York-based industrial PPE and construction equipment supplier with a 25-year reputation serving B2B contractors and direct buyers nationwide. They partnered with us for a 6-month engagement after their 2,000+ SKU Google Merchant Center feed faced widespread disapprovals, leaving their digital presence completely invisible.

We overhauled their product feed infrastructure, rebuilt their Google Ads strategy, and segmented B2B from B2C traffic. Over six months, this unlocked 4.6M impressions, generated $250K+ in attributed revenue, and delivered a 900% ROAS.

Case Study: The Argument
The idea that shaped the account

Most Shopping campaigns fail before the auction.

When Shopping underperforms, the instinct is to look at bids, budgets and audiences. But a product that fails feed validation never enters the auction at all. No bid strategy reaches it. No budget increase finds it. It simply isn't there, and the account reports it as normal.

Optimising the campaign

Tuning an account that can't serve.

Bids get adjusted, budgets get raised, audiences get refined. The products causing the problem are disapproved, so none of it reaches them. Spend concentrates on the handful of SKUs that happened to pass validation, and the catalog's real range never gets tested.

  • SKUs eligible A fraction
  • Catalog coverage Partial
  • Impressions available Capped
  • Ceiling Structural
Fixing the feed first

Make the catalog eligible, then compete.

Identifiers completed, shipping configuration corrected, compliance errors resolved across every SKU. Only once the full catalog could serve did campaign structure and bidding start to matter, and then they mattered enormously.

  • SKUs eligible 2,000+
  • Impressions earned 4.6M
  • Average CPC $1.14
  • Return on ad spend 900%
Case Study: Starting Position
Starting position

Twenty-five years offline. No shelf space online.

PowerPak Civil & Safety wholesales construction safety and PPE equipment from Congers, New York, serving contractors and direct buyers across the US. The offline business was established and trusted. The digital one was invisible.

Not invisible in the sense of ranking poorly. Invisible in the sense that a large share of the catalog was not permitted to appear in Google Shopping at all.

Four problems sat underneath that.


  • The feed wouldn't validate

    Merchant Center errors were causing product disapprovals across a 2,000-SKU catalog. Disapproved products don't bid, don't serve, and don't appear in performance reports as a problem. They simply produce nothing.


  • The catalog had no structure Google could read

    Two thousand SKUs with no product-type taxonomy and no segmentation. Even approved, they'd have competed against each other for the same budget with no way to bid differently on a $12 pair of gloves and a $400 fall-arrest harness.


  • Search, Shopping and Display shared one undifferentiated account

    No separation between B2B contractors buying in bulk and consumers buying single items, two buyers with different order values, different search language and different repeat behaviour, funded from the same pot.


  • Conversion tracking couldn't measure the outcome

    Without reliable tracking there was no basis for optimisation. Every decision would have been a guess dressed as data.

Case Study: The Build
The build

Four steps, and the first one took the longest.

The order matters. Nothing downstream of the feed can be evaluated until the feed is clean.

Step 01

Rebuild the product feed

Restructured and cleaned the full 2,000+ SKU feed. Resolved compliance errors, corrected shipping zone configuration, and worked the catalog until every product was approved and eligible to serve.

Unglamorous, slow, and the reason everything after it worked.

Step 02

Build a structure the catalog could scale into

Search, Shopping, Display and Brand campaigns organised separately, with explicit segmentation between B2B and B2C buyers, and seasonal and regional demand patterns built into the architecture rather than bolted on later.

Two thousand SKUs stopped competing against each other for one budget.

Step 03

Find the intent nobody was bidding on

Keyword and competitor research across the industrial PPE space, expanding into high-intent long-tail terms where a contractor searching a specific product specification is much closer to purchase than one searching a category.

Qualified traffic at $1.14 a click in a category most people assume is expensive.

Step 04

Optimise against real conversion data

Tracking rebuilt so performance could be measured properly, then continuous refinement of bids, placements and audience segments, with dynamic remarketing bringing back buyers who had viewed products without ordering.

900% return on ad spend across the six-month period.

Case Study: Campaign Breakdown
Campaign architecture

Four campaigns, four jobs.

Each budgeted and measured separately, so a campaign that worked could be scaled without one that didn't quietly absorbing the spend.

Search

High-intent B2B and industrial queries, contractors searching a specific product or specification, not browsing a category.

Shopping

Product-level visibility across the full approved catalog. The main driver of purchase-ready traffic once the feed was clean.

Display & remarketing

Reach beyond search volume, plus dynamic remarketing to bring back buyers who viewed products without ordering.

Brand

Defending searches for PowerPak's own name. Twenty-five years of reputation generates branded demand that competitors will happily intercept.

Feed management

Ongoing monitoring so approvals hold. Feeds decay: prices change, stock moves, requirements shift.

Conversion tracking

Rebuilt so revenue could be attributed to campaign and product, making optimisation a decision rather than a guess.

Case Study: What It Produced
What it produced

The arithmetic, in full.

Over six months the account generated 23,000 clicks from 4.6 million impressions at an average cost of $1.14. Those clicks converted at 1.5%, producing 357 orders and more than $250,000 in attributed revenue, a return of roughly nine times the spend.

The click-through rate looks modest at half a percent, which is what a Shopping and Display-weighted account produces. That's the correct trade in a category like this: broad product-level exposure across a large catalog, converted by a small number of high-value orders averaging around $700 each.

The number worth sitting with is the impressions. 4.6 million, from a catalog that, six months earlier, was largely not permitted to appear.

  • Impressions 4,600,000+
  • Clicks 23,000+
  • Average CPC $1.14
  • Conversion rate 1.5%
  • Orders 357
  • Attributed revenue $250,000+
Case Study: What Transfers
What transfers

If your Shopping campaigns underperform, check this first.

Open Diagnostics before Campaigns

Merchant Center Diagnostics tells you how many products are actually eligible. If that number is well below your catalog size, no bidding change will help.

Disapprovals are silent

A disapproved product doesn't error in your ad account. It produces nothing, and the report looks like ordinary underperformance.

Large catalogs need taxonomy

Two thousand SKUs in one undifferentiated campaign means you cannot bid differently on a $12 item and a $400 one. Product types and custom labels are what give you that control.

Case Study: Services Used
Case Study: More Work
Case Study: Final CTA
Next step

How many of your products can actually appear?

Most merchants have never compared their eligible product count against their catalog size. Give us read access to Merchant Center and Google Ads and we'll tell you what's disapproved, why, and what it's costing you. Findings are yours to keep either way.

Read-only access · No obligation · You keep your accounts