You have to know the past to understand the present - Case Study Patient.co.uk - SISTRIX

Most migration post mortems are stories about redirects. This one is not, and that is precisely why it is worth studying. Patient executed a domain change from patient.co.uk to patient.info in June 2015 for a stated strategic reason: the organisation wanted a top level international domain to serve what it described as an increasingly global audience. That is a rational business objective. The SEO outcome was still catastrophic, because the migration changed something that redirects cannot carry across, which is what the brand looks like to the people typing the query.

Sistrix's framing in the original piece, that you have to know the past to understand the present, is the operative point. The visibility chart alone tells you a site fell over. Only the history explains why it never got back up.

The numbers

Figures as reported by Sistrix. Visibility Index is a weighted measure of ranking positions against a fixed keyword set, so it is comparable across time for one domain but is not a traffic count.

Cause one: the brand, not the redirect

A domain migration is a request to the search engine to transfer the accumulated signals of one hostname to another. Redirects handle the machine readable half of that. They do nothing about the half that lives in people's heads.

Patient.co.uk had spent two decades becoming the thing a British person typed when they wanted to check a symptom without going to the GP. The .co.uk suffix was load bearing in that identity: it signalled local, established and institutional in a vertical where those signals decide whether a user clicks. The .info suffix carries close to the opposite association for a UK audience, and Sistrix's evidence for the rejection was direct rather than theoretical: Google Suggest kept showing patient uk as a leading search variant long after the migration, which means users were still trying to navigate to a brand that no longer existed at that address.

The practitioner test before a rebrand is cheap. Pull the branded query set from Search Console and check how much of it contains a geographic or TLD cue, then ask whether the new name preserves that cue.

# Search Console API: how much demand is tied to the old brand string?
# Filter Performance by query containing your legacy brand token,
# then compare against total clicks.

dimensions: ["query"]
dimensionFilterGroups:
  filters:
    dimension: QUERY
    operator: CONTAINS
    expression: "patient uk"

# If legacy branded demand is a double digit share of clicks,
# a name change is a demand risk, not a technical one.

Cause two: forcing recovery with links

Sistrix's second finding is the one most likely to be repeated by anyone reading this. Patient.info acquired more backlinks in roughly two and a half years than patient.co.uk had earned in twenty, with anchor text that was tightly targeted rather than naturally varied.

Read that as a response to panic. Visibility falls, the obvious lever is authority, so budget goes into acquisition. But link velocity that outstrips a brand's actual public prominence is one of the clearest patterns a spam system can learn, and concentrated commercial anchors make it clearer still. The intervention intended to fix the drop plausibly deepened it.

The comparison is a general practitioner framework applied to the characteristics Sistrix reported, not a claim about specific link sources.

Cause three: the slot does not stay open

The third finding is the strategic one. Sistrix reported that Healthline.com ended up taking the UK position Patient had held. This is what makes health and finance migrations different in kind rather than degree from a migration in a low competition niche.

When a dominant domain in a your money or your life vertical vacates the top of the results, the traffic does not go unserved. It flows to whoever is next, and that competitor then accumulates its own engagement history, its own link acquisition off the back of the new position, and its own quality rater exposure. Recovery is not a matter of restoring your former signals, because the benchmark has moved. You now have to beat an incumbent that is measurably stronger than it was when you left.

What a defensible version of this migration would have looked like

The right hand column is standard migration practice, not hindsight invented for this write up.

The checks worth stealing from this case

Three of these are technical and two are not, which is roughly the correct ratio for a domain change.

# 1. Every legacy URL resolves in exactly one hop to a 200
while read u; do
  printf '%s ' "$u"
  curl -s -o /dev/null -w '%{num_redirects} %{http_code} %{redirect_url}\n' -L "$u"
done < legacy_urls.txt | awk '$2 != 1 || $3 != 200'

# 2. Legacy domain still resolves and its certificate is valid
curl -sI https://patient.co.uk/ | head -1

# 3. Branded demand share, from Search Console, before you commit

The two non technical checks: measure how much of your click volume arrives on branded queries that contain the element you are about to discard, and identify who ranks immediately below you in your top thirty commercial terms, because that is who inherits the traffic if the migration slips.

Related reading on this site

The Clarks international consolidation is the technical counterpart to this case, where the strategy was sound and the redirects were not. The Home Depot Canada geo blocking case shows a third failure mode, access rather than architecture or brand, and the migrations case study index collects the set.

FAQ

Source: https://www.sistrix.com/blog/you-have-to-know-the-past-to-understand-the-present-case-study-patient-co-uk/