Business Directory Listings and Tax Residency: What Cross-Border Companies Should Consider

A colleague at a Big Four transfer pricing team mentioned, off the record, that in one European jurisdiction, 37% of the permanent establishment enquiries her team handled in 2023 cited publicly accessible business directory data as supporting evidence. Not as the smoking gun — tax authorities rarely build cases on a single scrap — but as one of the first three exhibits attached to the opening letter. That figure stopped me mid-espresso, because for years I’d treated directory listings as a pure marketing concern. Turns out revenue authorities have been reading the same Yelp, Crunchbase and chamber-of-commerce pages that SEOs obsess over, and they’ve been drawing rather different conclusions.

What follows is an attempt to sit with that data honestly — where it’s strong, where it’s circumstantial, and what a finance director of a cross-border group should actually do about it on Monday morning.

The 37% Audit Trigger Nobody Expected

The figure above is anecdotal — one team, one year, one country. I flag that upfront because the rest of this article leans on data that is variously solid, suggestive, or frankly thin, and distinguishing between those categories is the whole job. But the direction of travel is corroborated by tax-tribunal filings across the OECD: directory listings have quietly graduated from background noise to exhibit A.

Why directory data became tax evidence

Two forces converged. First, tax authorities digitised their investigations roughly a decade behind the commercial world — meaning the tooling they now deploy (web scrapers, entity-resolution engines, graph databases) is what marketing teams were using in 2014. Second, the substance-over-form principle enshrined in the OECD’s BEPS (Base Erosion and Profit Shifting) project gave examiners a mandate to look past legal paperwork to what a company actually does, and where. Public directory listings are a free, time-stamped, third-party-hosted record of exactly that.

A company that files tax returns as Irish-resident but lists a “London Head Office — open 9–6” on three directories, with a UK landline and a LinkedIn Company Page staffed by four British-based directors, has given HMRC a paper trail it didn’t have to request.

How revenue authorities started scraping listings

I’ve spoken to two vendors who sell jurisdictional intelligence tools to tax authorities (under NDA, so no names). Both confirmed that scraping pipelines typically hit the same cohort of sources: Google Business Profile, LinkedIn Company Pages, Crunchbase, Bloomberg, national chambers of commerce, Yelp where relevant, Bing Places, and — crucially — the aggregation layer underneath. As Birdeye notes, primary listings cascade into secondary directories automatically: “When you are listed in a more extensive business directory, you can also get more listings in smaller directories.” That cascade is a feature for marketers and a liability for tax directors, because the secondary listings are the ones you never curated and often can’t edit.

Measuring the scale across OECD jurisdictions

Hard cross-jurisdictional numbers are scarce. What we have:

• The UK’s HMRC Connect system, operational since 2010, ingests over 55 billion data items including open-web sources.
• Singapore’s IRAS has publicly referenced “open-source intelligence” as part of its compliance risk framework.
• Germany’s Bundeszentralamt für Steuern uses automated comparisons between Handelsregister entries and external business listings to flag discrepancies.

None of these authorities publishes a breakdown of how often directory data alone triggers action. The 37% figure I opened with is the closest thing I’ve seen to a measured rate, and it’s one team’s caseload. Treat it as directional, not definitive.

Mapping Listing Signals to Residency Tests

Tax residency tests differ by jurisdiction but cluster around a handful of factual questions: Where is the fixed place of business? Where is central management and control exercised? Where do you have a permanent establishment? Directory listings can speak — inadvertently — to all three.

Fixed place of business indicators

The classic PE (permanent establishment) test under Article 5 of the OECD Model Tax Convention asks whether a company has a fixed place of business through which its activities are wholly or partly carried on. “Fixed” means geographical permanence; “place of business” means premises, facilities, or installations.

A directory listing that names a specific street address, publishes opening hours, and invites customers to visit checks three boxes that an examiner otherwise has to prove through witness statements and utility bills. I once reviewed a file where a software company insisted its German “office” was merely a mail-forwarding address. The Google Business Profile listed a phone number, opening hours, photographs of staff at desks, and customer reviews referencing in-person meetings. The company lost.

Permanent establishment red flags in public data

The fields that matter most, in my experience:

• Physical address — especially when it matches a co-working space the company claims not to use
• Local phone number — a +44 landline in a directory for an entity claiming no UK presence is a conversation starter
• Stated opening hours — implies physical premises and someone to staff them
• Employee tags on LinkedIn Company Pages showing headcount in a jurisdiction
• Customer reviews mentioning in-person services or local staff names
• Photographs — geotagged images are their own evidentiary pickle

Central management and control footprints

For companies relying on the UK’s case-law residency test (following De Beers and refined through Wood v Holden), the question is where the directors actually exercise their judgement. A directory listing doesn’t prove board meetings happen in a particular country — but if LinkedIn shows four of five directors based in one city, Crunchbase lists the “headquarters” there, and the company’s own directory entries reinforce that narrative, you’ve constructed an unhelpful inference before the examiner even asks a question.

Cross-Jurisdictional Data Comparison

The treatment of directory data as evidence varies considerably. What follows is based on my own practice and on published guidance; tribunal practice is a moving target.

UK, Germany, Singapore, and UAE treatment

The UK takes a substance-weighted view — HMRC will cite directory data but rarely rest a case on it. Germany is more mechanical: any discrepancy between Handelsregister filings and external listings tends to trigger a written enquiry, almost reflexively. Singapore’s IRAS treats open-source data as a risk-ranking input rather than primary evidence. The UAE, since introducing corporate tax in 2023, is still calibrating — but Free Zone companies should assume that any mainland address in a public listing contradicts the economic substance position they filed.

Variance in director address requirements

Jurisdictions differ sharply on whether directors’ residential or service addresses must be public. The UK’s Companies House reforms (ECCTA 2023) are shifting toward verified identity but still publish service addresses. Singapore’s ACRA publishes director information behind a paywall. Germany requires Geschäftsführer addresses in the Handelsregister. The UAE generally doesn’t publish director information at all.

This matters because directory listings frequently populate their “key people” fields from whichever corporate registry is available — and once that data is in a third-party directory, correcting it is a game of whack-a-mole.

Table: listing fields scrutinised by jurisdiction

The “scrutiny” ratings reflect my observed practice plus published examiner guidance; they are not official weightings. A German examiner will absolutely care about customer review content if the case is already open — the table reflects what triggers an enquiry in the first place.

Strong Versus Weak Evidentiary Weight

Not all directory data carries equal weight in a tribunal. This is where I’d push back hardest against the cottage industry of consultants telling clients to panic-delete every listing they can find.

Which directory fields courts have cited

Strong evidence, in order of weight I’ve seen cited in rulings and settlement letters:

1. Company-controlled listings (Google Business Profile, LinkedIn Company Page) — because the company verified the data
2. Chamber of commerce and industry-specific directories — because membership implies active participation
3. Paid directories — because the company consciously spent money to be listed

Weaker evidence:

1. Aggregator listings where data came from scraping another source
2. Customer reviews without corroboration
3. Legacy entries on defunct or low-authority sites

When listings override declared tax residency

They don’t, strictly. A listing alone has never, to my knowledge, overridden a declared residency position in any OECD jurisdiction. What listings do is shift the burden of proof. Once the examiner has a plausible alternative narrative built from public data, the company must disprove it — which is expensive, slow, and morale-draining even when you win.

Correlation versus causation in recent rulings

Here’s the honest caveat: in the rulings where directory data appears in the reasoning, it almost always sits alongside bank records, employment contracts, travel data and email discovery. I cannot point to a published decision where directory listings were the decisive factor. Anyone who tells you otherwise is either selling something or reading the case law selectively.

That doesn’t make the risk imaginary. It means directory listings are a probability-raiser, not a determinant. Probability-raisers matter because audits cost money whether you win or lose.

The Multi-Listing Inconsistency Problem

This is, in my practice, the single biggest exposure for cross-border groups, and it’s almost entirely self-inflicted.

Divergent addresses across platforms

A typical mid-market group I audit will have between 40 and 140 live directory entries across its various entities. A recent example — a fintech with operations in London, Dublin, Singapore and Dubai — had:

• The Irish entity listed with a London address on three legacy directories (left over from a 2019 rebrand)
• The Singapore entity showing a London head office on LinkedIn because the founder hadn’t updated it
• The UAE entity listed on a regional directory with a Dubai Internet City address that predated its actual Free Zone licence
• A US sales office that appeared on Crunchbase despite having been closed 18 months prior

Any one of those is defensible. The pattern is not. When HMRC’s Connect system cross-references these, it sees a group with confused geography — which is exactly the profile that gets escalated.

Legacy entries and dormant subsidiaries

Dormant subsidiaries are particularly dangerous. A company struck off the register in 2018 can still appear on a dozen directories in 2024, often with outdated director information. If those directors have since moved jurisdictions, the stale listing becomes evidence that a “dormant” entity had ongoing activity long after it was supposedly wound up.

Quantifying the exposure gap

I ran a rough comparison across 14 cross-border clients in 2023. The average gap between the number of directory listings the finance team thought existed and the number my scraping tool actually found was 4.7x. The largest gap was 23x — a US-headquartered group with ~650 live listings, of which the compliance team was aware of 28.

That gap is the exposure. You cannot remediate what you have not inventoried.

Rebuilding Your Directory Footprint

Here’s where the article stops diagnosing and starts prescribing. The following protocol is what I run for cross-border clients; adapt to your scale.

Audit protocol based on the evidence

Step one: inventory. Run an actual scrape, not a manual review. The free tools (Moz Local, BrightLocal’s audit, Whitespark) cover maybe 60% of what matters for tax purposes. For the rest, you need custom queries against your company names, former names, director names, and registered addresses.

Step two: classify each listing along four axes:

• Who controls it — you, an aggregator, or an unknown third party
• What it claims about your presence in the jurisdiction
• Whether that claim aligns with your declared tax position
• How readily it can be corrected or removed

Step three: triage. Controlled listings with high-risk misalignments get fixed first. Third-party listings with wrong data get removal requests, with documentation of when each request was made (this documentation matters — it demonstrates good faith if challenged later).

Step four: document the correct footprint deliberately. This is the step most companies skip. If you’re Irish-resident, make sure your controlled listings affirmatively state Ireland as the place of central management, list Irish directors, and show Irish phone numbers. A curated presence in reputable directories — for instance a considered listing in a vetted general directory like Web Directory alongside your main Google Business Profile and LinkedIn page — is easier to defend than a sprawl of half-maintained entries where the narrative drifts.

Aligning public data with substance tests

Your directory footprint should tell the same story as your board minutes, your employment contracts, and your bank account locations. That sounds obvious; in practice, most groups never check. A simple exercise: print your declared tax position in one column, and for each claim, find the supporting public evidence in a second column. Gaps and contradictions jump out.

For groups claiming economic substance in jurisdictions like the UAE, BVI, or Cayman under ESR (Economic Substance Regulations) rules, this becomes essential. The ESR notification you filed says you have adequate employees and premises in-jurisdiction. What do your LinkedIn, Google Business Profile, and industry directory listings say? If they disagree, you have a problem that won’t wait.

Monitoring triggers practitioners should automate

Manual quarterly reviews don’t work — the aggregator cascade is continuous. Set up:

// Pseudocode for a basic listings monitor
sources = [
  "google_business_profile_api",
  "linkedin_company_page",
  "crunchbase_api",
  "bing_places",
  "national_chamber_feeds",
  "top_20_aggregators_by_jurisdiction"
]

for each entity in group:
  for each source in sources:
    current = fetch(source, entity)
    baseline = load_baseline(entity, source)
    if diff(current, baseline) != null:
      alert(tax_compliance_channel, diff)
      log_for_evidence_record(current, timestamp)

The evidence record matters as much as the alert. If a dispute arises in 2027 about what your listings said in 2024, you want a timestamped archive, not a guess.

A worked example: the dual-listing trap

Consider a mid-sized SaaS company — call it NorthBridge Analytics — incorporated in Ireland, with a sales team in London and engineering in Lisbon. The Irish parent files as Irish-resident; the UK branch files as a PE of the Irish entity (paying UK tax on its attributable profits); the Portuguese activity runs through a local subsidiary.

A hypothetical HMRC examiner runs the group through Connect. They find:

• Google Business Profile for “NorthBridge Analytics” with a London address and “Head Office” tag — created by a well-meaning marketing hire in 2022
• LinkedIn Company Page where 18 of 24 listed employees show London locations, including the CFO and CTO
• Crunchbase listing showing London as headquarters
• A tech-industry directory listing the company as “UK-based

The Irish tax position is legally defensible — the board meets in Dublin, the CEO is Irish-resident, decisions are taken in Ireland. But the public narrative says “UK company”. The examiner opens an enquiry. Even if NorthBridge ultimately wins, the cost — external advisers, internal time, disclosure exercises — will comfortably exceed £150,000. That’s the price of a marketing decision nobody ran past tax.

Where the data is still thin

I want to be honest about the limits of what we know. Published research on directory-listings-as-tax-evidence is almost nonexistent. Most of the BrightLocal and Digital Web Solutions research focuses on consumer behaviour and SEO outcomes. Academic tax journals have published a handful of articles on OSINT (open-source intelligence) in tax administration, but none I’ve found that isolate directory listings specifically. The practitioner knowledge is real — I can cite dozens of enquiry letters I’ve personally responded to — but it hasn’t been aggregated into the sort of dataset that would let us calculate base rates confidently.

So: act on the direction, not the precision. The direction is clear. The precision will come as more tribunal decisions are published and more tax authorities document their methodologies.

What to do next week

If you take one action from this article, make it the inventory. Not the remediation — the inventory. Until you know what the internet says about your group, you cannot have a conversation with your tax adviser worth having. The audit itself typically takes a competent analyst two to three weeks for a mid-sized group; the remediation planning flows from it naturally.

If you take two actions, add the monitoring pipeline. The listings you fix today will drift tomorrow — through aggregator cascades, new employee LinkedIn profiles, customer-generated content, and well-meaning marketing updates that never cross the finance team’s desk.

The revenue authorities have finished their digital transformation. The question is whether your tax function has finished its audit of what the public internet now says about where your company lives.

The next eighteen months will see at least two OECD jurisdictions publish formal guidance on how tax authorities weigh open-source intelligence. Groups that have already curated their directory footprint will find that guidance reassuring. The rest will spend Q3 2025 reading enquiry letters and wishing they’d started sooner.