What this playbook covers and when to run it
The playbook covers the whole life of every tool the marketing team pays for, uses or has left running: how a need becomes a request, how a request becomes a decision, how a new tool is brought in safely, how access and licences are kept honest, how each renewal is decided before its notice date, how a tool leaves, and how the stack as a whole is audited and redesigned. It does not teach any of those stages. The marketing technology pillar explains what the stack is for and the questions that keep it lean, the stack hub sets out the layers and why ownership matters, and each phase below hands off to the guide or procedure that owns the method.
Run it when any of these is true: renewals arrive as invoices rather than decisions; nobody can say how many tools the team has; tools appear on payment cards without anyone outside the team knowing; a former colleague or agency still holds a login; or two teams pay separately for the same job. Run it also before a stack redesign, because a redesign without a running programme slides back into the old shape within a year.
Like the other playbooks in the playbooks library, it is a starting shape. A team with a handful of tools and one marketer may run it in an afternoon a month. A large team with many suppliers and integrations will split it across marketing operations, finance and engineering. The phases stay the same; the frequency and the number of people change.
Roles and responsibilities
- Stack owner. Usually marketing operations. Owns the stack record, the request route and log, the renewal calendar and the access register, and runs every review in this playbook. Does not own every tool.
- Tool owners. One named person in the team that uses each tool. Answers for its value, its configuration, its users and its success measure, and brings the evidence to each renewal decision.
- Requesters and team leads. Raise needs through the one route, answer its questions and say what the tool must achieve. Team leads confirm the need is real before an evaluation starts.
- Finance or procurement partner. Holds contracts, notice dates, purchase orders and card spend, and confirms who may commit money. The spending authority guide covers payment cards and platform budgets as well as contracts.
- Privacy and security contact. Whoever owns data protection. Reviews what personal data a tool will receive before it arrives, and approves retention and deletion when a tool leaves.
- Engineering partner. Involved whenever a tool needs a script on the site, an integration or a change to where data is collected.
A RACI matrix is a useful way to write these roles down per phase, because the commonest failure is a gate where everyone is consulted and nobody decides. In a small team one person may hold the stack owner, finance and privacy roles at once; the responsibilities still need to be visible, because the unowned one is the one that slips.
Cadence, inputs and outputs
The rhythm below is a starting point. Adjust it to the number of tools, how much personal data they hold and how often they change, not to the calendar you would like to have.
| Cadence | Owner | Inputs | Outputs |
|---|---|---|---|
| Weekly | Stack owner | New requests, trials with an end date, card sign-ups found by finance | Triage outcomes in the request log; evaluations opened with a decision date |
| Per tool | Stack owner, tool owner, privacy contact | The decision record from the evaluation | An onboarded tool, an adoption review and a stack record marked live |
| Monthly | Stack owner, finance partner | The renewal calendar built from notice dates | Renewal reviews started in time; notice given where a tool is leaving |
| Quarterly | Stack owner, tool owners | User lists, licence terms, the request log | Access removed or downgraded, seats right-sized, repeated needs sent to the roadmap |
| Annually | Stack owner, team leads, finance, engineering | The stack record, the use-case map, the year's logs | A keep, fix, consolidate or retire decision on every tool and an updated roadmap |
Each cadence feeds the next. The weekly log shows the quarterly review which needs keep returning; the quarterly reviews give the renewal decisions their evidence; the renewal decisions and the logs give the yearly audit a head start. Tool costs and savings belong in the monthly numbers too: the reporting cadence playbook sets who sees which figures when, and the guide to tracking spend against budget notes that annual subscriptions may be booked as prepayments and that software is sometimes booked outside marketing's own cost lines, so ask finance to look beyond them when the stack record is checked against spend.
The playbook in phases
Take stock: inventory, owners and the renewal calendar
Nothing else in this playbook works without a record of what is there. The stack audit tutorial finds every tool from invoices, card statements, logins and the live site, and records each one's owner, cost, renewal date, users and connections. The first audit becomes the stack record that every later phase reads and updates. Alongside it, the integration register writes down what connects to what, including the manual exports nobody thinks of as integrations, and who holds the credentials. Finish this phase by building the renewal calendar from notice dates rather than renewal dates, as the renewal review tutorial insists, because the notice date is the one that can be missed.
Open one route in and triage requests
Every new tool, free tier, trial and module arrives through one route, which the guide to handling requests for new tools designs: a short set of questions about the job to be done, the users, the data, the owner and the budget, and a triage against what the business already owns. Each request leaves triage with one of four outcomes: already covered, extend what exists, evaluate, or decline with a reason. The categories guide helps when a product describes itself in unfamiliar terms; it uses the marketing technology landscape published by chiefmartec.com as vocabulary, not as a shopping list. Requests for AI tools take the extra questions in the guide to evaluating AI tools. The stack owner triages weekly so that waiting never becomes a reason to sign up on a card.
Evaluate and decide with a written record
A request that reaches "evaluate" becomes a selection with requirements written before any demonstration, a proof of concept on real use cases and a decision record that says what was chosen, against which alternatives, on which evidence and with which known gaps. The guide to choosing and evaluating tools sets out the method, and the marketing tool evaluation template holds the record. A weighted scoring matrix keeps the comparison honest when several people score. Where the choice is between one suite and several specialist tools, the suite versus best-of-breed comparison decides layer by layer. The decision record's success measure is the one the renewal review will read later, so write it now.
Onboard the tool
A tool that has been chosen enters through the marketing tool onboarding SOP: approval, a data and privacy review, access and roles, integration and tagging, testing the data flow, training and an adoption review some weeks after launch. The SOP requires the data protection lead to approve before any Personal DataAny information relating to an identified or identifiable living person, such as a name, email address, location data or an online identifier. reaches the tool, and checks whether consent or another lawful basis covers the use, as the GDPR requires. Three guides support it: test environments and test data keep trials out of live reports, the consent map is updated when a new tool reads or holds consent state from the CMPSoftware that collects, stores and passes on visitors' consent choices for cookies and tracking., and any script goes through tag governance with an entry in the tag inventory. The stack record changes from onboarding to live only after the adoption review.
Review access and seats
Quarterly, as a starting rhythm, match every account in every tool to a current person, supplier or integration, confirm each tool has an owner and a deputy administrator, and replace shared logins with named accounts. The access review tutorial runs this in order and ends with triggers for joiners, movers and leavers between reviews; the handover tutorial for leavers covers the work that leaves with a person. In the same sitting, the licence and seat review compares what each licence bills on with what is used: seats against active users, contact counts held in the CRMThe system that stores an organisation's contacts, companies, deals and interactions, and the processes built around it. or email platform against what the programme needs, modules nobody opened and duplicate subscriptions. Changes that only take effect at renewal go into the renewal calendar with their notice dates.
Decide each renewal before its notice date
The monthly read of the renewal calendar starts each review early enough that every outcome is still open. The renewal review tutorial gathers the decision record, measures use, reads the success measure against its baseline, lists incidents and workarounds, checks what has changed in data handling and terms, and re-scores the tool. The outcome is one of four: renew, renegotiate, replace or retire. The tool owner brings the evidence; the budget holder and the stack owner decide; the finance partner records the outcome against the contract. A review that cannot finish before the notice date is escalated, not allowed to renew by default.
Retire tools without loose ends
A tool leaving the stack is retired in the order the retirement tutorial gives: confirm the decision and who signs it off, list every dependency, decide where each job goes, read the contract for notice, export and deletion terms, export what must be kept with consent and suppression records intact, switch off integrations and remove tags, remove users and credentials, give notice and request deletion with written confirmation. The retention guide decides what is kept and for how long, and makes deletion reach every copy. Update the stack record, the decision record and the integration register on the same day, or the next audit finds a ghost.
Audit the stack and set the roadmap
Once a year, as a starting rhythm, rerun the stack audit from the stack record rather than from scratch, using the year's request log, access findings, renewal outcomes and the tag and analytics audits it already relies on. Every tool leaves the audit with a decision: keep, fix, consolidate or retire, each with an owner and a date. Gaps and repeated requests go to the roadmap that the guide to designing a marketing stack sequences in phases, usually starting from the data layer, whether that is a Data WarehouseA database designed for analysis that stores data loaded from many systems, such as analytics, ad platforms and the CRM, so it can be joined and queried. or a CDPUnified first-party customer profiles.. Bring the roadmap's next phase to the quarterly planning playbook, so stack work competes for time and budget with everything else rather than happening in the gaps.
The stack record: one register and the logs around it
The programme depends on a small set of records kept in one place, owned by the stack owner and readable by every tool owner. Most of them already exist in the guides; the playbook's job is to keep them together and current. Of these, the stack recordThe live register of every marketing tool: owner, deputy, cost, notice date, users, use cases and status, kept current by onboarding, reviews and retirement. is the one every phase of the programme reads and updates.
| Record | What it holds | Updated when | Guide |
|---|---|---|---|
| Stack record | Every tool with owner, deputy, cost, notice date, users, use cases and status | Any phase changes a tool | Stack audit |
| Request log | Every request with its answers, triage outcome, decision owner and follow-up | Weekly triage | Handling requests |
| Decision records | What was chosen, why, the success measure and the review date | Each evaluation and each renewal | Choosing and evaluating tools |
| Integration register | Each connection, its fields, identifier, schedule, credentials and check | Any integration change | Documenting integrations |
| Access register | Every account matched to a person, supplier or integration | Quarterly review and every joiner, mover or leaver | Access review |
| Renewal calendar | Notice dates, review start dates and owners | Monthly read and every contract change | Renewal review |
Two rules keep the records honest. A change that is not reflected in the stack record is treated as not done, which is the same rule the stack design guide applies to its roadmap artefacts. And the stack record lives where the team already works, with a link from the team's onboarding pack, because a register that only one person can find will be one person's opinion.
An insurance broker's first year of stack management
Brissoway Insurance Brokers, an invented home and motor insurance broker with a marketing team of five, starts the playbook after two renewals arrive as invoices in the same month. The first audit finds 23 tools, 4 of them bought on cards, 2 survey tools doing the same job and 3 accounts still held by a former agency. The stack owner opens one request route and triages on Mondays; 9 requests arrive in the first quarter, 5 of them answered "already covered". The renewal calendar shows 7 notice dates in the year; the review of a call recording tool ends in a renegotiation to 6 seats from 10. One survey tool is retired with its consent records exported and deletion confirmed in writing. The first quarterly access review removes 11 accounts and replaces 2 shared logins. At the year-end audit the stack holds 19 tools, every one with an owner and a deputy, and the roadmap's first phase goes into the next quarter's plan.
Measuring the playbook
The playbook works when tools enter on purpose, earn their renewals and leave cleanly. Operating signals show whether the rhythm is holding: whether triage happens on its scheduled day, whether every renewal review starts before its notice date, whether each quarterly access and seat review is completed and recorded, and how many tools are found at the yearly audit that the stack record did not know about. That last number is the best single test of the programme; it should shrink each year.
Outcome signals show whether the stack is getting better: the share of tools with a named owner and deputy, the number of renewals decided rather than defaulted, tools retired with written deletion confirmation, and repeated requests that reached the roadmap rather than a card. Report them in the team's existing monthly and quarterly packs rather than in a separate stack report, so tool decisions sit beside the results they are meant to support.
Common mistakes
- Auditing once and stopping. A first audit without a running programme is a snapshot. Keep the stack record current through every phase, or the next audit starts from scratch.
- Several routes in. If requests arrive by chat, email, meetings and card statements, the process sees only the ones people choose to send. Make one route the only route and answer it quickly.
- Working from renewal dates. The notice date comes first. A review that starts when the invoice arrives can only renew.
- Owners who cannot log in. A tool owner who cannot administer, renew or answer for the tool is a name on a list. Confirm every owner and deputy at each access reviewA periodic check of who can use each marketing tool: every account matched to a current person, supplier or integration, with owners, deputies and narrow roles..
- Skipping the privacy review for trials. A trial that receives real customer data is a real use of personal data. Put it through the onboarding procedure first.
- Cancelling instead of retiring. Stopping a contract leaves tags, integrations, users and data behind. Retire in order and confirm deletion in writing.
- Stack work that never reaches the plan. Roadmap phases left out of the quarterly plan never get time. Bring them to planning like any other piece of work.
