Tools and providers
Part of How to choose programmatic advertising tools and providers
A scoring rubric makes programmatic advertising software selection defensible
How UK agencies can run programmatic advertising software selection with a weighted rubric covering data rights, auction mechanics and supplier scoring.
- Open a dated decision record with the purchase question, people involved and campaign scope.
- Separate offers by buying model and record who is responsible for delivery.
- Agree weighted criteria and evidence rules before supplier demonstrations.
- Give shortlisted suppliers the same brief and assess their written replies.
- Test the leading option in a paid pilot against the rubric.
- Keep dated score versions and revisit them when terms change.
Put the decision on record
Start with a dated record because it fixes what the team is deciding and lets later reviewers trace the choice to the material available at the time. Name the buyer, budget owner, client media in scope, intended campaign brief, reviewers and date.
Keep the original brief beside supplier replies and scored documents. Name the person who maintains the record, so later reviewers can understand the basis of the decision and find the evidence used. This trail is useful when an answer is unclear or the team changes its view after assessment.
Separate unlike offers
Sort candidates by buying model before scoring. A self-serve demand-side platform licence provides software access, while a managed service includes delivery responsibilities. Keeping unlike offers apart avoids treating different accountabilities as equivalent and makes the comparison meaningful for the purchase being considered.
Self-serve seat or managed service?
What is the job to be done?
Self-serve seat -> bid controls and audience tools
Managed service -> service levels and exit route
List each supplier's deliverables and give every responsibility a named owner. Treat a managed contract as a delivery arrangement rather than software access. If a provider offers both models, score each separately unless the purchase uses the same arrangement.
Fix the marks before meetings
Choose the criteria and their relative importance before supplier demonstrations begin. Agree the weights in advance, use one scoring grid and keep evidence beside every mark. A selection checklist can organise the material, while a product comparison can show findings consistently.
| Criterion | Weight | What a 5 looks like |
|---|---|---|
| Data rights and privacy | 5 | Identifies controller and processor roles, lawful basis, any PECR consent requirements, data use, retention and subprocessors; supports the claims with relevant contract terms and evidence. |
| Auction transparency | 5 | Explains bidding, auction mechanics, supply path and deductions clearly enough for the buyer to understand how spend and outcomes are reported. |
| Fees and commercial terms | 3 | Separately discloses media-spend commission, CPM technology fees, minimum spend and managed-service fees where applicable, with the charging basis in writing. |
| Reporting and data portability | 5 | Provides itemised reporting and usable exports, including media, technology and agency fees, so the buyer can retain and reproduce reports outside the platform. |
| Inventory and supply controls | 3 | Shows the sources of inventory and how the buyer can apply controls, exclusions and suitability settings. |
| Campaign controls and optimisation | 3 | Explains available bid and campaign controls, what the system optimises towards and how changes can be reviewed. |
| Creative and personalisation | 1 | Documents supported formats and personalisation logic, including the data used and the buyer controls available. |
| Exit and termination | 5 | Sets out termination rights, data return or export, deletion arrangements and any continuing charges in clear contractual terms. |
Use 5 for a critical criterion, 3 for an important one and 1 for a useful differentiator. These values are starting points, not a universal answer. Record the agreed weights, score each criterion from 1 to 5, multiply each weight by its score and total the marks. Attach written evidence to every score.
Request comparable proof
Send each shortlisted supplier the same written campaign brief, fixed budget, audience definition and response deadline. Judge the replies against the agreed criteria rather than the polish of separate presentations. Where an answer is missing or unclear, lower the score for that requirement instead of accepting an unsupported claim.
Questions that expose weak claims
- How does bidding actually work?
- How is creative personalised?
- What happens on termination?
- Who owns the reporting data?
Candidates could include Google Display & Video 360 (DV360), The Trade Desk, Amazon DSP and Adform. These are examples to assess, not endorsements. Where practical, hide supplier names during written review and compare replies side by side. A supplier comparison can help reveal differences in fees or missing reporting detail.
For privacy scoring, request written detail on controller and processor roles, lawful basis, PECR consent requirements, data use, retention and subprocessors. Consult the ICO's electronic and telephone marketing guidance where relevant. General assurances should not receive a strong mark without evidence to support them.
Enter zero if a supplier will not provide a written answer, rather than leaving the cell blank. Keep the common brief and response deadline with the marks. A later reviewer can then check that candidates had an equivalent opportunity to reply and that missing information was not overlooked during moderation.
Check performance in delivery
Use a paid pilot to test the leading option against the original rubric, rather than relying on a free trial. Live delivery can reveal inventory, fees and reporting quality that a demonstration cannot establish. Reuse the agreed criteria to assess the same claims as in the initial comparison.
For example, a team paying £1,500 a month for a pilot seat can inspect fee lines, inventory reports and export files. Score the same criteria again using the pilot evidence, keep that material with the updated marks and note whether the order of candidates has changed.
Reporting quality and support responsiveness may alter the ranking. If the result moves, reassess the evidence instead of preserving the earlier outcome by default. A revised score is useful when it records what the pilot established and explains why the team changed its view.
Keep a traceable version history
Keep a dated copy of each rubric version and put a review date in the contract calendar. Fees or data terms may change, making an earlier ranking out of date. Retain both previous and current scores so reviewers can see what changed and which evidence led to the update.
Six to eight criteria is a workable range: fewer may omit data rights or exit terms, while a longer list can slow assessment and make reviewers less consistent. Price need not carry the greatest weight by itself. Check commercial contracts alongside headline charges, since unclear fees can make leaving costly.
Include the media buyer, the person accountable for client data and someone from finance in the review. Reviewers working from written evidence can produce a steadier result than relying on one senior opinion. When a fee model or data term changes, update the relevant marks and preserve earlier versions rather than overwriting the decision history.
