How to run employer branding reviews without guesswork

What to take away

Why reviews differ from pitches

A pitch is a performance. A review is a record. The two rarely agree, and the gap widens as the contract value grows.

Most employer branding decisions in England are made by a small internal group: a talent acquisition lead, a procurement officer, and sometimes a communications manager. None of them has time to test every claim.

That is why the method matters more than the verdict. A structured review forces every vendor to answer the same questions.

Start with your own definition of the role. The CIPD's overview of the people profession sets out the scope of HR and people work, which helps you decide whether employer branding sits with talent acquisition, internal communications or a standalone team. That decision changes what you are buying.

Once the scope is fixed, write the questions down. A vendor that resists written questions before a meeting is telling you something useful.

A shortlist built on pitch performance alone tends to favour the best presenter. References then reveal a different ranking, and the whole exercise repeats a year later.

Written answers also create a paper trail. When two finalists score within a point of each other, the written record is what lets you justify one over the other to finance.

Two scorers, the same questions and a dated file will survive an audit. Anecdote and instinct will not.

The evidence you can check

Some claims are verifiable. Use those first, because they cost nothing and remove weak candidates quickly.

Company registration and filing history are public. Search the provider on Companies House to confirm the legal entity, its incorporation date and whether accounts are overdue. A trading name that does not match the registered company is a question, not necessarily a problem.

Advertising claims are checkable too. The Advertising Standards Authority publishes its adjudications, and the ASA rulings library is the fastest way to see whether a provider or its client has been pulled up for a misleading recruitment claim. Read the reasoning, not just the outcome.

The labour market context is public. The ONS employment and employee types data shows how many people are employed, self-employed or on temporary contracts, which is the population your employer brand has to reach.

Platform capability is usually documented by the vendor. LinkedIn publishes a comparison of its hiring products covering Jobs, Recruiter and Recruiter Lite, which is enough to check whether a quoted feature set matches the tier being sold.

Everything else, from culture fit to account team quality, is judgement. Score it, but label it as judgement.

Work in ascending order of cost. Registry checks come first, published rulings second, references third, and only then site visits or paid discovery. Each stage should be able to eliminate a provider.

Note the date of every check. A solvency statement from March says little about a contract signed in September.

Where a provider belongs to a group, check the parent's filing history as well as the trading company's. Group accounts can show financial pressure that a tidy subsidiary page hides.

Keep a PDF or screenshot of every page you check. Pages change, and a dated copy settles later arguments about what a vendor claimed.

Building the scoring rubric

A rubric is a table with weights. The weights should reflect your own risk, not the vendor's strengths.

Use a 1 to 5 scale for each criterion. Multiply by the weight to get a weighted score. Keep the total at 100 so the result reads as a percentage.

Score each provider twice: once on the written submission, once after the pitch. Where the two scores diverge by more than a point on any criterion, ask why.

An illustrative example: a team paying £400 a month for a recruitment marketing platform might weight cost transparency at 10, while a team spending £40,000 on a rebrand might weight track record and exit at 20 each. The weights follow the money.

Test the rubric on a provider you have already ruled out. If the scores do not reflect that decision, the weights are wrong.

Judgement criteria need anchors. Write one sentence describing a 1 and another describing a 5 for each subjective row, so two scorers read culture fit the same way.

Keep the table to eight or nine rows. Longer rubrics produce fatigue, and the last criteria get scored carelessly.

Sector matters. A care provider, a bank and a logistics firm carry different regulatory risk, so compliance and data handling deserve more weight in regulated settings than in a consumer brand.

Platform versus agency

These are different purchases and should not share a rubric without adjustment.

A platform is judged on capability, data terms, integration and price per seat or per job. A shortlist of tools is best compared in a single table with identical columns. The employer branding product comparison in England walks through how to line up features without letting sales tiers distort the picture.

Volume hiring, multilingual careers sites and always-on advertising point to a platform. A narrative problem, a values reset or a reputation issue points to an agency. Buying the wrong category costs more than picking the wrong supplier inside the right one.

Ask each platform vendor to price your own volumes, not a template. Ask each agency to name the people who would work on the account and how many days a month that means. The answers rarely arrive in the same format, which is why the rubric has to carry the comparison.

Where the two overlap, for example a platform sold with a managed service, split the score. Give the software 60 per cent of the weight and the service 40, or whatever ratio matches your risk.

Do not let a strong demo carry a weak contract. The demo is a sample of the best day; the contract records the worst.

An agency is judged on people, process and evidence of past work. The employer branding agency reviews in England covers how to read a case study critically, including what a client reference can and cannot tell you.

Running the review process

Set a timetable and publish it to the shortlist. Providers behave better when the deadline is real.

Steps 3 and 5 are where most processes fail. Scoring after the pitch, with no pre-pitch baseline, turns the rubric into a memory test.

Keep the sessions equal in length and format. If one provider gets ninety minutes and another gets forty, the scores stop being comparable.

Allow written questions from the shortlist and share every answer with all of them. Selective briefing creates an advantage that resurfaces later as a delivery dispute.

Build in one site visit or working session per finalist if the contract is large. Watching a team solve a live problem tells you more than a rehearsal does.

The employer branding review methodology in England sets out the full sequence in more detail, including how to handle a provider that refuses to answer written questions.

Due diligence on providers

Due diligence is not a single check. It is a short list of checks you can evidence.

Legal entity: confirm the registered name and number. If the entity is newly incorporated, ask who carries the liability for prior work.

Financial standing: read the latest accounts filed at Companies House. Small companies may file abbreviated accounts; that limits what you can see, so ask directly for a statement of solvency if the contract value is material.

Insurance: professional indemnity and public liability, with certificates dated within the last twelve months.

Data protection: a written description of what personal data is processed, the lawful basis, retention periods and the breach notification route. If the provider cannot describe this in plain English, that is a finding.

References: two clients of similar size in a similar sector, contacted by phone. Ask what went wrong, not what went well. The answer to the second question is usually more informative.

Compliance history: check whether the provider or its clients appear in published ASA rulings. One ruling is not disqualifying. A pattern is.

Ask who will actually do the work. Agencies subcontract design, video and media buying more often than their pitch suggests. Name every subcontractor in the file and note which of them will touch candidate data.

Confirm who owns the assets at the end. Logos, film rushes, careers site code and audience lists belong in the contract, not in an email thread.

The employer branding selection checklist in England turns these checks into a one-page form you can attach to the procurement file.

Comparing costs fairly

Total cost of ownership beats headline price. Add the annual licence, implementation, training, content production, and the internal time to manage the relationship.

Ask for pricing in the same units. Per seat, per job, per project and per day are not comparable until you convert them to a twelve-month figure for your own volumes.

Watch for three common distortions.

Put every provider's twelve-month figure in one column. If a provider will not give one, score cost transparency at 1 and say so in the file.

Separate media spend from fees. Paid social and job board budgets can dwarf the service fee, and mixing them into one line makes the comparison meaningless.

Ask what happens at renewal. An uplift clause tied to an inflation index, or a notice period longer than the contract term, can wipe out the saving that won the business.

Budget for the internal cost too. A marketing manager spending a day a week on agency liaison is a real cost, even when it never appears on an invoice.

Governance and record keeping

A review that is not recorded did not happen. Keep one file per procurement with the rubric, the scores, the evidence and the final decision.

Set a review date for the contract itself. Annual is common; for a fast-moving platform, six months is reasonable. The point is to compare performance against the same criteria you used to buy.

Assign an owner. A named person, not a committee, should hold the file and the review date.

Agree the triggers for an early review in writing. A change of ownership, a new account lead, two missed reporting cycles or a data breach should reopen the file without argument.

This is also where operational planning begins. The employer branding operations and delivery guide for 2027 covers how to move from a signed contract to a working delivery rhythm, including the handover documents to request at kick-off.

At the review, score the provider on the same rubric and compare the two sets of numbers. Movement on one criterion matters more than the overall total.

Common questions

How many providers should we score?

Three to five is workable. Fewer than three gives you no comparison; more than five consumes time without changing the ranking. Score all of them on the same rubric.

Can we use a vendor's own case studies as evidence?

Yes, but label them as vendor-supplied. A case study supports a claim; it does not verify it. Verify by phoning the named client and asking what went wrong.

What if a provider refuses to answer written questions?

Score the relevant criteria at 1 and note the refusal in the file. Some providers reserve detail for the pitch stage, which is reasonable, but a blanket refusal is a risk signal.

How often should the rubric change?

Review the weights before each procurement, not during. Changing weights mid-process invalidates the comparison and is hard to defend if the decision is challenged.