Edinburgh Glasgow analytics clusters: how buyers differ

What to take away

Two clusters: Edinburgh fintech and public sector versus Glasgow agencies

Edinburgh and Glasgow sit forty-odd miles apart and buy measurement in almost opposite ways. Edinburgh's demand comes from financial services, data platforms and public bodies. Glasgow's comes from agencies and the higher education institutions that feed them staff.

Edinburgh fintech analytics teams are usually inside a regulated firm. Their measurement question is rarely "which channel worked" and more often "can we prove what we did with this person's data". Consent records, retention schedules and processor contracts sit in the same conversation as campaign reporting.

Edinburgh public-sector analytics faces a different constraint again. Scottish Government publications shape how public bodies describe outcomes, and those descriptions feed into how measurement is commissioned. Reporting tends to be aggregated, published and slow to change.

Glasgow digital agencies are the opposite: fast, client-facing, and judged on whether the monthly report explains a spike. Many are small enough that measurement is one person's job alongside paid media buying.

That difference matters commercially. If you are selling measurement services into Scotland, you are selling two products: assurance to Edinburgh, throughput to Glasgow.

ONS employment and labour market data shows how concentrated both labour pools are in professional and digital roles, which is why hiring for measurement skills across the Edinburgh and Glasgow corridor is competitive rather than plentiful.

Higher education in Glasgow also supplies graduates who arrive knowing analytics platforms but not UK GDPR case law. Agencies train that gap on the job, which is why their consent practice often lags their reporting practice.

Consent handling differences across the two cities

Consent handling in Edinburgh starts with the record. Fintech teams tend to keep consent state in a customer platform and pass a signal to marketing tools, so the marketing stack never holds the source of truth.

That architecture is expensive and slow to change, which is the point. It means a withdrawal propagates everywhere at once instead of being reconciled across five dashboards.

Glasgow digital agencies more often take consent from a consent management platform bolted onto the website, then rely on the ad platforms to respect it. It works until a client asks for evidence of what was shown to whom. PECR places that consent duty on the organisation running the site, not on the ad platform.

Public-sector analytics in Edinburgh leans on a third model: collect less, explain more. Where a lawful basis other than consent applies under the UK GDPR and the Data Protection Act 2018, the emphasis moves to transparency notices and data protection impact assessments.

ICO guidance on privacy by design is the practical reference here, because it pushes teams to build the privacy decision into the product rather than the policy page. Both cities would benefit from reading it earlier in a build.

Where the two clusters converge is on the awkward cases: logged-in users, app traffic and offline conversions. Those are the points where consent state is hardest to keep accurate, and where Edinburgh and Glasgow teams most often ask for help.

A first-party data strategy is the usual fix, because it moves the identity question back to a place the organisation controls. Teams that have done that work find consent arguments get shorter, not longer.

Data clean rooms in Scottish measurement practice

Data clean rooms in Scotland are still a minority practice, and the pattern of adoption is telling. They appear where two organisations hold overlapping audiences and neither will hand over identifiers.

That describes a bank and a broadcaster, a university and a public body, or a fintech and a retail partner. It rarely describes a Glasgow agency and its client, because the agency usually holds both sides of the data already.

The commercial case for a clean room is strongest when the match itself is the insight: how many of our customers saw your campaign, and what did they do next. In Edinburgh and Glasgow that question tends to arrive from the fintech side.

The practical barriers are mundane. Clean room vendors scale their work to the volume of data being matched, Scottish campaigns are often small, and a match rate that looks respectable in London can look thin against a Scottish audience of a few hundred thousand.

There is also a skills constraint. Running a clean room means writing queries against a schema you do not own, and that is a different skill from building a dashboard. Glasgow agencies report hiring difficulty here more often than Edinburgh firms do.

Our supplier comparison is a useful counterpoint. It shows how much of a clean room engagement sits in the managed service around the technology rather than in the technology itself.

For most Scottish buyers the sensible first step is a single, narrow match with a partner you already trust, measured against a holdout. If that produces a decision, scale it. If it produces a chart, stop.

Attribution approaches compared

Attribution in Edinburgh fintech tends to be conservative and modelled. Long consideration cycles, regulated products and limited conversion volume make last-click useless, so teams blend marketing mix modelling with incrementality tests.

Edinburgh public-sector analytics rarely uses attribution at all in the commercial sense. The question is reach and outcome against a published target, not credit allocation between channels.

Glasgow digital agencies live in platform attribution: conversion APIs, modelled conversions and dashboards built inside the ad platforms. It is fast, cheap and increasingly opaque as platform modelling fills gaps left by signal loss.

That opacity is the commercial opening. When a client cannot explain why reported conversions exceed actual sales, the agency that can reconcile the two wins the account.

The reconciliation usually needs an incrementality test, and the design questions are the same ones that come up in any incrementality testing programme: what is the control, how long does it run, and who agreed the decision rule in advance.

Scottish campaigns add a wrinkle. Audiences are smaller, so tests need longer to reach significance, and geo holdouts are harder to construct without spilling into neighbouring areas.

A workable compromise for most Scottish teams is a switchback test on a single channel, run long enough to cover a full purchase cycle, with the result written into the next planning round rather than the next invoice.

What buyers should test before choosing a cluster

Before you choose between an Edinburgh supplier and a Glasgow one, test the parts of the engagement that fail quietly. Dashboards rarely fail. Deletion, consent state and reconciliation do.

Work through three tests before you shortlist:

Worked example: reconciling reported and actual conversions

Take a campaign with a known sales figure and make each shortlisted supplier explain the gap themselves.

A short due diligence process will separate the clusters faster than any pitch. Vendor due diligence that runs from contracting party through to verified deletion is the part Scottish buyers most often skip, and it is the part that costs most to fix later. It should also include a check of the ICO's register of fee payers, which is public and searchable.

Agencies should also check the IPA knowledge centre for agency-side guidance on measurement and media practice, because it sets expectations about what a credible agency measurement offer looks like.

Finally, decide what you are buying. Edinburgh suppliers sell assurance and defensible method. Glasgow suppliers sell speed and interpretation. Neither is wrong, but a brief that asks for both from one supplier usually gets neither.

Common questions