Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Templates

Marketing Tool Evaluation Template

A marketing tool evaluation template: requirements, weighted criteria, vendor scores with evidence, trial findings, integration checks, total cost and decision.
Templateintermediate10 min readUpdated

What this template is for

Use the workbook whenever a tool will hold customer data, connect to other systems or cost enough that someone outside marketing will ask why it was chosen. That covers email and messaging platforms, customer relationship management systems, a CDP, analytics and reporting tools, experimentation and personalisation tools, and the consent layer. A browser extension one person uses for a week does not need one. A tool already in the stack being stretched to a new use case does, in a lighter form, because the question is the same: does it do this job with our data? For an AI tool or feature, the guide to evaluating AI tools lists the extra criteria to add to the workbook, from output testing on your own tasks to model changes and exit. For an email platform, the guide to choosing an email platform supplies the requirement rows, from signing mail with the business's own domain to suppression that survives a re-import and a complete export.

The workbook sits between two other documents. Before it, the stack design supplies the use cases and the layer the tool must serve, which is where the requirements come from. After it, the marketing tool onboarding SOP takes the decision and turns it into a configured, integrated, reviewed tool. The workbook is neither: it is the record of the selection, and it is the document the renewal conversation starts from. If the category vocabulary is new, the MarTech categories guide explains what each kind of tool does; the chiefmartec.com landscape listed in the sources shows how many products compete in each, which is a reason to fix the criteria first rather than a menu to order from.

Keep the workbook with the contract once the decision is made. A year later, when the person who ran the evaluation has moved on, it is the only place that explains why the tool was chosen, which gaps were known and what the renewal should check.

The template

Copy the table into a document, or into a spreadsheet with one tab per field where the scoring needs columns. Fill the first three rows before any vendor demo is booked.

FieldWhat to write
Problem and requirementsThe problem in one sentence, the use cases, and each requirement marked essential, important or useful, grouped as functional, integration, data and operational
Stack fitThe layer the tool serves, the systems it connects to, what it replaces and what it overlaps with
Criteria and weightsThe scoring criteria and the weight of each, agreed and dated before the first demo
Vendor scores and evidenceEach candidate, including the incumbent, scored per criterion with the evidence and its source
Trial findingsThe use cases tested, the data used, the success definition written in advance, and what happened for each
Integration and data checksInbound and outbound connections, identifier, consent handling, where data is processed, retention, deletion and export, each answered yes, no or workaround
Total costLicence, implementation, integration, training, owner time, data and usage costs, workarounds and exit cost over the contract term and a renewal
Decision recordThe choice, the reasons, the known gaps, any dissent, the onboarding owner and the review date

How to fill each section

Problem and requirements

Write the problem as the business feels it, then the use cases the tool must serve, then the requirements. A requirement is a sentence that can be checked: "send a pre-arrival email built from the booking record on the day the balance is paid" is a requirement; "good automation" is not. Group requirements as functional, integration, data and governance, and operational, as the tool selection guide recommends, and mark each essential, important or useful. Resist marking everything essential, because a list where everything is essential cannot rank anything. Ask the people who will run, report on, integrate and govern the tool to add their requirements now, since a requirement added after the demo is one the demo was designed to avoid.

Stack fit

Name the layer the tool serves — collection, data, decisioning, activation or measurement — and draw the connections it needs. The stack design guide explains the layers and why most tools span more than one. Write down what the tool replaces, so the old tool's retirement is part of the plan, and what it overlaps with, so two tools do not end up doing the same job. If the tool is meant to hold the customer record, say whether it will read from an existing source of truth or become one, and which first-party data it will store.

Criteria and weights

Turn the requirements into scoring criteria — use-case fit, integration, identity and consent, data handling, usability for the owner, reporting, governance, vendor, total cost and exit are the usual set — and weight them from the stack design. Integration, identity and data handling tend to carry the heaviest weights because they cannot be fixed after purchase. Date the weights. If they change after a demo, the workbook should show the change and the reason, because that is exactly where a selection is steered. The weighted scoring matrix framework covers setting weights before scoring and testing whether the ranking survives a change of weights.

Vendor scores and evidence

Score every candidate on every criterion, and score the incumbent too, because an evaluation that scores only challengers compares a known tool's faults with unknown tools' promises. Beside each score, write the evidence and where it came from: the trial, the documentation, a reference call, the contract. A score whose only evidence is the demo should say so. Where weighted totals are close, essential requirements decide; a tool that fails an essential requirement does not win on the sum of the others.

Trial findings

Choose two or three use cases for difficulty, not ease: the hardest integration, the one that needs identity across sources, the one the owner will run most often. Give every candidate the same data and the same success definition, written before the trial starts, and a fixed period. Record what worked, what needed the vendor's help, what needed a workaround and what failed. A vendor that cannot support a trial with real data is itself a finding, and it belongs in the score.

Integration and data checks

Answer each check from documentation and a short test. What does the tool receive, from where, how fast? What does it send back to analytics and the data layer? Which identifier does it key on, and do the events it will receive carry that identifier? Can it read and act on consent per purpose, and does its tag work with the CMP and tag manager? Where is data processed and stored, by which sub-processors, for how long, and how is deletion confirmed? The data flow and integrations guide explains each question. Where the tool will process personal data, the data protection lead should read the vendor's terms during the evaluation; the GDPR, listed in the sources, sets out the obligations, and this template is general direction rather than legal advice. Every "workaround" answer is a cost to carry into the next row.

Total cost

Cost every candidate on the same basis over the contract term and one renewal: licence, implementation, integration time, training, the owner's time to run it, usage and data charges as contacts or events grow, the workarounds the checks found and the cost of leaving. Include the incumbent, whose implementation is already paid. Note the contract terms that change the cost later: price at renewal and any cap on increases, notice periods, usage tiers, and export and deletion on exit.

Decision record

Write what was chosen, against which alternatives, and why, in a paragraph a successor can read. List the known gaps, any dissent from the people involved, the onboarding owner who will run the onboarding SOP, and the review date with the measures it will check. A decision record that names its gaps is not a weakness; it is what lets the renewal be judged on evidence. The tutorial on reviewing a marketing tool before it renews reads this record at the end of the term and re-scores the tool on the same criteria as any alternative.

Filled example

Below, the workbook is completed in summary for an invented business. The vendors are anonymised and the figures and names are illustrative.

Illustrative example — not a real result

Wychcombe Holiday Cottages: choosing an email and CRM tool

FieldWychcombe's workbook
Problem and requirementsGuest and owner records live in three spreadsheets and a basic newsletter service, so repeat guests get the same generic newsletter as first-time enquirers and nobody can see a guest's past stays. 19 requirements, 7 essential, including: import bookings nightly from the booking platform with stay dates and property; send a pre-arrival email on balance payment; suppress guests with a stay in progress from promotions; record marketing consent per purpose with date and source; export all contacts and history on exit
Stack fitActivation and a light customer record, sitting between the booking platform (system of record for stays and payments) and GA4. Replaces the newsletter service and the three spreadsheets. Overlaps with the booking platform's own guest emails, which stay for transactional messages only. The booking platform stays the source of truth for stays; the new tool holds marketing consent and engagement as first-party data
Criteria and weightsAgreed 3 March, before demos: use-case fit 3, integration with the booking platform 3, consent and data handling 3, usability for a two-person team 2, total cost 3, reporting 1, vendor and support 2, exit 2
Vendor scores and evidenceVendor A 84, Vendor B 71, Vendor C 77, incumbent newsletter service 41 (weighted, each criterion scored 1 to 5, maximum 95). A met all 7 essentials in the trial; B could not import stay dates as dates without a paid connector (documentation and trial); C met all essentials but its support answered in 3 to 5 days during the trial
Trial findingsTwo weeks, same anonymised export of 2,400 guest records for each vendor. Use case 1, pre-arrival email on balance payment: A and C sent correctly; B needed a manual date field. Use case 2, suppress guests mid-stay: A worked from the nightly import; C worked only with an hourly sync on a higher tier. Success definitions written 10 March
Integration and data checksA: native connector to the booking platform (yes); sends opens and clicks to GA4 with UTM tags (yes); keys on email address, which every booking carries (yes); consent per purpose with timestamp (yes); data hosted in the EU with a published sub-processor list (yes); full export as CSV and via API (yes). One workaround: owners' newsletters need a separate list, maintained by hand
Total costOver 3 years: A £9,600 licence, £1,800 set-up help, about 40 hours of internal time; C £8,100 licence but £2,700 extra for the hourly-sync tier; incumbent £2,200 plus the spreadsheet time it does not save
Decision recordVendor A chosen on 31 March for meeting all essentials with the least integration work. Known gap: no SMS, deferred to next year. Dissent: the bookings manager preferred C's interface; recorded, and usability will be re-checked at review. Onboarding owner: Martha Quill, Marketing Manager. Review in 6 months: repeat-booking enquiries from the guest list, time spent on newsletters, and whether the owners' list workaround is still manual

At the six-month review the owners' list workaround was still manual but took under an hour a month, and the team kept it rather than paying for a second audience feature.

Common mistakes

  • The demo as the requirement. Requirements written after a demo describe the product just seen. Write them, and agree the weights, before any vendor is contacted.
  • Everything essential. A list with no priorities cannot separate candidates. Mark essentials sparingly and let them decide close calls.
  • Scores without evidence. A number from a demo is an impression. Record where each score came from and prefer the trial and the contract over the pitch.
  • Leaving out the incumbent. Without it, the comparison is between known faults and promised features. Score the current tool on the same criteria.
  • Integration and data checked last. These are the criteria that cannot be fixed after signing. Run the checks, with the privacy owner, during the evaluation.
  • Licence as total cost. Implementation, integration, owner time, usage growth and workarounds often outweigh the licence. Cost every candidate on the same basis over the term and a renewal.
  • No exit plan. A tool without a usable export holds the data hostage at renewal. Test the export during the trial.
  • No decision record. Without one, the renewal is argued from memory. Record the reasons, the gaps, the dissent and the review date.

The other planning and decision documents are in the templates library.

Terms used on this page

Frameworks

Where this page is used

Also related

Sources

  1. Marketing technology landscape — chiefmartec.com
  2. Regulation (EU) 2016/679 (GDPR) — EUR-Lex

Related