GroundWork SEO research · September 2026
Low visibility. Clear opportunities.
A practical read on what is broken, what is promising, and what to do in what order.
Sample report — anonymised from a real GroundWork engagement, September 2026. Client, competitor and legacy hostnames are replaced; every number is as researched.
The short version
The current search footprint is tiny, but the research found a credible route out of it. Most of the first moves are specific and measurable.
- 0commercial pages ranking under their own URL
- 1,450monthly searches in the leading UK opportunity cluster
- 36published pages still carrying Webflow template content
- 0citations across the AI search engines checked
What the research says
-
The commercial site is effectively invisible
The only two rankings found are still attributed to legacy blog URLs. Neither sits in the top 20, and the service site ranks for nothing in either market.
-
The technical base is decent, with a messy public edge
The crawl score is 92.1, but 36 template pages, inherited team profiles, missing canonical signals and a three-hop legacy redirect chain are getting in the way.
-
Offshore development is the clearest opening
The cluster carries 1,450 monthly UK searches. The strongest term has difficulty 0, the checked results are weak, and the consultancy has real delivery locations to talk about.
-
Managed cloud connects organic and AI search
"What is managed cloud?" combines 480 traditional searches, 156 AI searches and difficulty 4. It also gives the thin cloud page a useful hub-and-spoke structure.
-
Dutch demand is worth testing properly
"Webapplicatie laten maken" offers 430 monthly searches at very low difficulty and meaningful commercial intent. It needs native Dutch copy, not a translation pass.
A sensible order of work
First 1-2 weeks
Remove the avoidable noise
- Delete the template leak Remove the 36 products, categories, checkout and invented careers pages.
- Fix the people pages Use real names, titles and descriptions, with redirects from inherited slugs.
- Check canonical output Resolve the crawler and direct-request mismatch in Search Console.
- Shorten legacy redirects Send old blog URLs straight to their final www destination in one hop.
Next 4-8 weeks
Build the pages with a job to do
- Offshore development centre Turn the existing article and delivery model into a proper service page.
- Managed cloud hub Rebuild the thin cloud page around migration, AWS, DevOps, cost and security.
- Dutch web application page Commission native commercial copy and use it as the market test.
- Internal linking Connect case studies, explainers and service pages around each cluster.
Once the foundations are live
Earn coverage and expand carefully
- Publish the cloud explainer Nest “What is managed cloud?” below the service hub for search and AI citation potential.
- Work the round-ups Start with Clutch and the third-party lists already appearing in the data.
- Measure before multiplying Use the offshore rewrite and Dutch page to decide which content pattern travels.
- Add the next clusters DevOps, AWS and website translation follow only when the first pages show traction.
The 5 reports
-
Site Explorer
Current rankings, traffic, top pages and the opportunity created by legacy blog attribution.
- Just 2 keywords rank across both markets, none in the top 20
- client-site ranks for nothing in either market
- A 1,600/mo Dutch term sits at position 77 on a page that already exists
-
Keyword Research
Seven viable page clusters across the UK and Netherlands, with build order and SERP checks.
- 7 pages the keyword set can support, UK and NL combined
- Top UK cluster: 1,450 searches/mo at zero difficulty
- Only 1 of 3 checked SERPs came back genuinely weak
-
Content Gap
Where competitors rank, what their pages contain, and which gaps are worth closing first.
- 20 keywords held by 2+ competitors — the client ranks for none
- 584 words on the client's cloud page vs. 3,117 on the page beating it
- competitor-c (a UK web and app agency) alone holds 300+ top-20 rankings on its full footprint
-
Technical Audit
A strong underlying crawl score with a short list of public-facing issues to clear up.
- 36 of 110 sitemap URLs still carry the Webflow template's title
- Zero canonical tags found anywhere on the site
- Infrastructure itself is sound — 92.1 on-page crawl score
-
AI Search Visibility
Why the client is not cited today, which formats get cited, and where organic and AI opportunity overlap.
- Zero client citations on Google AI Overview or ChatGPT
- Competitor C alone holds 805 citations on Google AI Overview
- "What is managed cloud" already draws 156 AI searches a month
Three reasonable ways to tackle it
These are options, not packages. The right one depends on how much the existing site helps or hinders the work.
-
Rebuild around the evidence
The broader route is a lean rebuild with a clearer service architecture, proper English and Dutch pathways, useful content foundations, and technical SEO handled from the start.
I would use frontier LLM models within my proven agentic build process to move quickly, while keeping the important editorial and technical checks human-led.
This becomes the cleaner choice if the current Webflow structure keeps turning simple changes into cleanup work.
-
Tidy and unblock
Keep the current site. Fix the template pages, people pages, canonicals and redirects, then watch whether the existing URLs consolidate.
-
Build the priority pages
Keep the platform, but add the offshore, cloud and Dutch pages with the structure, copy and internal links the research points to.
My read: start with the technical tidy-up either way. If the site stays easy to work with, add the priority pages there. If it keeps fighting the plan, rebuild once and give the new structure a proper run.