Selected work

Five problems. Five different answers. None of them "spend more".

Each of these starts with what was actually wrong, which is usually not what the client thought was wrong. Numbers are from client analytics and CRM systems, not platform-reported figures.

A note on naming

Client names below are withheld or anonymised at their request. If you are building your own site from this template, replace these with named studies as soon as you have permission — named case studies convert roughly twice as well as anonymous ones.

D2C · Apparel

6.4× ROAS by fixing the feed, not the bids

The actual problem

A profitable Shopping account had plateaued and the previous agency's answer was to raise bids. The actual constraint was the product feed: 41% of SKUs were disapproved, mis-categorised or missing GTINs, so nearly half the catalogue could not serve at all.

What we changed

Rebuilt the product feed with complete attributes, correct taxonomy and GTINs
Added custom labels for margin tiers so Performance Max could be split by profitability
Excluded loss-making SKUs from advertising entirely
Set conversion values to gross margin rather than order value
Added brand exclusions so PMax stopped absorbing branded search

The lesson

Spend went down. The account was never bid-constrained; it was inventory-constrained by its own feed.

Services used: Google Ads · Ecommerce Management

Figures from the client's own analytics and CRM. Comparison period: the six months before engagement versus months 7–12.
Metric	Before	After
Blended ROAS	2.1×	6.4×
SKUs actively serving	59%	98%
Cost per acquisition	$62	$27
Monthly revenue	$184k	$511k
Ad spend	$88k	$80k

B2B SaaS

+312% organic clicks without a single link-building campaign

The actual problem

812 indexed pages, most of them thin programmatic variations, competing with each other for the same queries. Non-brand organic had been flat for two years despite consistent publishing.

What we changed

Consolidated 812 pages into 140 genuinely useful ones with 301 redirects
Rebuilt internal linking around six topical clusters
Added 40–60 word answer blocks beneath question-formatted headings on every commercial page
Deployed complete JSON-LD across all templates and published llms.txt
Established a quarterly refresh cycle on the top 30 pages

The lesson

Deleting 83% of the site tripled its traffic. Thin content is not neutral — it actively suppresses the pages you want ranking.

Services used: SEO Services · Generative Engine Optimization · Content Marketing

Figures from the client's own analytics and CRM. Comparison period: the six months before engagement versus months 7–12.
Metric	Before	After
Non-brand clicks / month	4,100	16,900
Indexed pages	812	140
Demo requests / month	31	118
AI Overview appearances	2	47
Avg. position, priority terms	18.4	6.1

Marketplace · Home & Living

27% of revenue moved to owned channels in five months

The actual problem

Entirely dependent on marketplace demand with no email list, no first-party data and no way to reach a past customer. Every sale had to be re-bought.

What we changed

Built WhatsApp opt-in at checkout and on packaging inserts with compliant consent capture
Implemented six email flows: welcome, cart, browse, post-purchase, replenishment, win-back
Deployed COD confirmation flow on WhatsApp, cutting failed deliveries
Segmented by RFM and suppressed unengaged contacts to protect deliverability
Fed purchase data back into paid targeting as exclusions and lookalike seeds

The lesson

The paid account barely changed. Blended CAC halved because a growing share of revenue no longer required acquisition at all.

Services used: Email Marketing Automation · WhatsApp Marketing Automation · Ecommerce Management

Figures from the client's own analytics and CRM. Comparison period: the six months before engagement versus months 7–12.
Metric	Before	After
Owned-channel revenue share	3%	27%
Repeat purchase rate	11%	34%
Blended CAC	$44	$19
Failed COD deliveries	18%	7%
Email list size	2,400	41,000

B2B Services

62% lower cost per opportunity by ignoring cost per lead

The actual problem

LinkedIn was producing leads at $61 and the board was pleased. Sales was not: fewer than one in twenty became an opportunity, and the pipeline had not moved in three quarters.

What we changed

Rebuilt conversion definitions around CRM opportunity stage, not form fill
Implemented offline conversion imports so LinkedIn optimised toward qualified pipeline
Replaced broad job-title targeting with a matched account list built from closed-won data
Moved cold campaigns from demo requests to an original benchmark report
Switched to thought-leader ads from two executive profiles

The lesson

Cost per lead nearly tripled and the programme became three times more profitable. Optimising the wrong metric is worse than not optimising.

Services used: LinkedIn Ads · Analytics & Tracking

Figures from the client's own analytics and CRM. Comparison period: the six months before engagement versus months 7–12.
Metric	Before	After
Cost per lead	$61	$174
Lead-to-opportunity rate	4.8%	31%
Cost per opportunity	$1,270	$482
Pipeline per $ spent	1.6×	4.4×
Sales cycle length	142 days	97 days

Healthcare · Multi-clinic

From 11 to 63 booked appointments a week on the same budget

The actual problem

Six clinic locations sharing one generic landing page, no local SEO, and a booking form that failed on mobile Safari. Paid traffic was arriving and quietly bouncing.

What we changed

Built location-specific pages with correct LocalBusiness schema and embedded booking
Claimed and optimised all six Google Business Profiles with review response protocols
Fixed the mobile booking form — the single largest contributor to the result
Split campaigns by location with radius targeting and location-specific ad copy
Added call tracking so phone bookings stopped being invisible

The lesson

A broken mobile form was costing more than the entire media budget. Always check the destination before optimising the source.

Services used: SEO Services · Website Development · Conversion Rate Optimization

Figures from the client's own analytics and CRM. Comparison period: the six months before engagement versus months 7–12.
Metric	Before	After
Booked appointments / week	11	63
Mobile form completion rate	22%	71%
Cost per booked appointment	$96	$31
Google Business Profile calls	40/mo	310/mo
Local pack visibility	1 of 6 clinics	6 of 6 clinics

What these have in common

Four patterns show up in almost every engagement that works.

The stated problem was not the real one

In four of five, the presenting complaint was a channel problem and the cause was measurement, catalogue or a broken page.

Something got removed

Pages deleted, SKUs excluded, campaigns paused. Subtraction produced more gain than addition in every case.

The metric changed first

Each result required agreeing a better success metric before any tactic changed. Optimising the wrong number faster does not help.

It took months, not weeks

Every result above is measured at six to twelve months. Anything faster than that in these categories would be luck.

Your situation is probably one of these.

Most accounts we audit have a version of one of these five problems. Forty-five minutes on a call is usually enough to tell which.

Book your free strategy call See pricing