Case study · Copenhagen, Denmark
Almost Nordic: from zero to 72 page-1 keywords in 10 weeks.
A recipe publisher founded in Copenhagen in 2012 had thousands of recipes but not one keyword on Google's first page. We rebuilt its recipes for search, wrote and published new ones, and replaced its slow page builder with a custom WordPress theme.
- Client
- Almost Nordic (almostnordic.com)
- Industry
- Food & recipe publishing
- Platform
- WordPress
- Timeline
- 15 Aug to 21 Oct 2026 (10 weeks)
- Status
- Ongoing: new recipes publish every day
Results
The numbers, before and after.
0→72keywords on Google page 1July vs October 20260→21keywords in the top 310→113keywords in the top 200→156recipes rebuilt or newly written85 rebuilt + 71 new
The challenge
Thousands of recipes, invisible to Google.
We started on 15 August 2026 by checking every one of the site's 453 posts. The content was there, but Google could barely see it.
- 0187 real recipes had no recipe card and no recipe schema, so Google didn't treat them as recipes at all.
- 0255 of those recipes had a broken main image from 2018. Without a valid image, Google won't show a recipe rich result.
- 03Most pages were drawn by the Elementor page builder: 64 stylesheets and 22 scripts on the homepage alone.
- 04The flagship cassava cake recipe (5,400 searches a month) ranked for zero cassava keywords.
- 05Every category page had the same title, empty pages after the last real page returned "OK", and the list of user accounts was public.
- 06Result: in the five months before we started, not a single keyword reached Google's first page.
The redesign
Drag to compare before and after.
BeforeAfter
Before: archived 6 June 2026. Old pages are shown from web-archive copies, which didn't keep all the old styles; content and layout are accurate. After: the live site, 21 October 2026.
What we did
Six steps in ten weeks.
- 1Week 1
Audit and a new recipe layout
We sorted all 453 posts into real recipes, unit-conversion pages and articles, then designed a new recipe layout with the client.
- Hero photo next to a sticky "On this page" menu
- A proper recipe card with ingredients, method and times
- Step-by-step pictures and a FAQ
- Flagship cassava cake page rebuilt first
- 2Week 2
Fixing the broken recipes
47 recipes were hidden behind the page builder, so normal edits never showed. We found a safe, reversible way to rebuild them.
- 54 rebuilt, 47 published in one day
- Backup before every change
- Every page checked automatically before publishing
- 3Weeks 2 to 8
Rewrites based on real ranking data
We picked the pages closest to page 1, studied who ranks for each search, and rewrote them so Google and AI answers can quote them.
- 85 recipes rebuilt, 80 fully rewritten
- Times and quantities checked against three sources, never guessed
- No fake ratings or nutrition
- 4From week 5
New recipes people search for
We found dishes the site didn't cover yet, with real search demand and a results page that could be won.
- 148 ideas checked, 71 written, 77 rejected on evidence
- Three new recipes a day, one per kitchen
- About 456,000 US searches a month across the new targets
- 5Week 5
Faster production
One article took 81 minutes, mostly waiting. We changed the workflow to research and write in batches with one review.
- Ten articles prepared at once
- Same safety checks, done in batches
- 6Weeks 6 to 7
Redesign and a custom WordPress theme
A new design system, "Saffron & Ink", approved by the client, built as a custom WordPress block theme with no page builder.
- Homepage, recipe, category and author pages redesigned
- One clean schema output for the whole site
- 478 posts moved off the page builder, 0 left
- 6 theme releases in one week
Google rankings
From page 2 and beyond to page 1.
Keywords the site ranks for in Google US, month by month. Work started on 15 August.
SEO, AEO & GEO
What changed under the hood.
| Feature | Before | After |
|---|---|---|
| Recipe schema (JSON-LD) | Missing on 87 recipes | Exactly one per recipe page, checked on the live URL |
| Recipe card | Plain lists in the text | Card with tick-off ingredients, numbered steps, times, print |
| "On this page" and "Jump to recipe" | None | On every rebuilt recipe |
| FAQ with FAQPage schema | None | On the full rewrites |
| Main recipe images | 55 broken, one watermarked | 109 new photos, also used in schema and social previews |
| Category page titles | Same title on every archive | Real names, e.g. "Filipino Kitchen" |
| Pages past the last real page | Returned OK with no posts | Return 404 |
| Public list of user accounts | 15 accounts visible | Locked (401) |
| Page builder | Elementor, 50+ stylesheets | Custom lightweight block theme |
| Titles and descriptions | Generic | Written for real search intent |
| "Last updated" date | None | Visible on the page and in schema |
- Homepage HTML size
219.6 KB→ 142.1 KB (-35%) - Stylesheets on the homepage
64→ 10 - Script files on the homepage
22→ 5 - Posts drawn by the page builder
56→ 0 (478 checked)
New pages
Pages that didn't exist before.
Flagship rewrite: Filipino Cassava Cake 
New recipe: Nihari, Pakistani kitchen 
Recipe page on mobile
Delivered
Ten weeks of work, in numbers.
- 453posts checked and sorted
- 85recipes rebuilt
- 71new recipes written
- 109new recipe photos
- 597step-by-step illustrations
- 437backups before changes
How we kept it safe
- A backup before every change, so every post can be restored.
- An automatic check before publishing: a page that fails is never published.
- Checked on the live page, not just in the admin.
- No invented data: times and quantities from three sources, no fake ratings.
- No URL changed without approval, so no rankings were lost.
- Results measured with a control group, so we don't over-claim.
What's next
- Fix the hosting timeouts that sometimes take the site offline.
- Keep measuring every month, with the control group.
- Add author profiles and bios for every kitchen.
- Keep publishing three new recipes a day.
Tools and tech
- WordPress
- Custom block theme (PHP, theme.json)
- Recipe schema (JSON-LD)
- FAQPage & BreadcrumbList
- Python automation
- DataForSEO
- Surfer SEO
- Cloudflare
FAQ
Questions about this project.
How fast did Almost Nordic see results?
In the first month of work, keywords in Google's top 20 went from 10 to 55. By October 2026, ten weeks in, the site had 72 keywords on page 1 and 21 in the top 3, after zero page-1 keywords in the five months before.
What was the biggest SEO problem?
87 real recipes had no recipe card and no recipe schema, and 55 of them had broken images. Google couldn't treat them as recipes, so they could never get rich results, no matter how good the text was.
Why replace Elementor with a custom theme?
The page builder loaded 64 stylesheets and 22 scripts on the homepage and hid content edits on many posts. A custom WordPress block theme made pages lighter (35% less HTML), easier to update and cleaner for search engines.
Is the traffic going up too?
Rankings rose strongly. Estimated traffic is a model, not real analytics, and it dipped in September and October, mostly on unit-conversion pages we didn't work on and during hosting outages. We report it as measured and keep tracking it.
Your site next?
Want results like Almost Nordic?
Tell us about your website. On a free 30-minute call we'll show you what's holding it back and what we'd fix first.