Marketing change management: a lightweight process for agencies and in-house teams

What is marketing change management?

Marketing change management is the routine that decides who can change what, who signs off on risky edits, where each change is recorded and when the team looks back at its effect. It covers paid media, the website, tracking and the CRM, because a drop in leads can start in any of them.

It is not the heavy change advisory board of IT. An agency that edits 20 accounts a week can't wait three days for a committee. The goal is narrower: when a client asks why a number moved, you answer with a date, a name and a reason in minutes, not after a day of digging through five tools.

Without a process, the same three problems come back again and again:

• Overlapping edits. Two people change the same campaign in the same week, and nobody can tell which edit moved the result.
• Silent changes. A developer ships a new form, a client pauses a campaign, a platform applies a recommendation, and nobody on the team hears about it.
• Lost context. The platform history shows that a budget changed, but not why, what the plan was or what result was expected.

A light process fixes all three with four steps and one shared record. The marketing change log guide covers the record itself; this article covers the workflow around it.

What does a lightweight change process look like?

Four steps: request, approve, log and review. Most changes skip the approval step, so the process costs a few seconds per edit and one short meeting per week.

1. Request. Whoever wants a change writes one line: what, where, why and the expected result. For example: raise the Search budget from €80 to €100 a day because the campaign is limited by budget; expect about 15% more clicks. A message in a shared channel or a ticket is enough.
2. Approve. Only changes in the high-risk tier (next section) need a second person. The approver checks timing and overlap: is anything else changing in the same campaign this week?
3. Log. When the change goes live, record the date, owner, platform, before and after values and the reason. Log it the same day; a log filled in on Friday from memory misses half the edits.
4. Review. Once a week, read the week's changes next to the main metrics. Mark each change as worked, didn't work or too early to tell, and decide what to roll back.

Keep requests and the log in the same place when you can. The marketing change log template gives you the columns to start with.

Which marketing changes need approval and which only need logging?

Sort changes by how much they can move results and how hard they are to undo. Approve the top tier and log everything.

Two rules make the tiers work. First, set the thresholds per client: a 20% budget change matters on a small account and may be routine on a large one. Second, treat tracking as high risk every time. A change to a conversion action or a tag can make results look worse or better with no real change in sales, and it distorts every comparison that follows.

Also log the changes your team did not make: client edits, platform automations, site releases, price changes and outages. They explain results just as often as your own edits. Before peak periods, add a freeze window on top of the tiers, as described in the Black Friday change freeze article.

How do Google Ads and Google Tag Manager support approvals?

The platforms control who can edit, but only Tag Manager 360 has a built-in approval step, and none of them records why a change was made.

Google Ads

Google Ads has five access levels: Email-only, Read-only, Billing, Standard and Admin. Standard users can edit campaigns; Admin users can also give access and change other users' access levels. There is no approval queue, so anyone with Standard access can push a budget change live. A practical setup: give Read-only access to people who only report, and keep Admin to a small group.

After the fact, Change history (in the Campaigns menu) shows changes from the past 2 years with the user, the date and details such as previous budget amounts, filtered by date range, change type and campaign. Google notes that some account-level settings and changes made by Google representatives may not be listed. Through the API, the change_event resource only covers the past 30 days. The Google Ads change history guide shows how to read it.

Google Tag Manager

Container permissions are Read, Edit, Approve and Publish. Edit can change a workspace but can't create versions or publish; Approve can create versions but can't publish; Publish has full rights. In Tag Manager 360, users with Edit access can request approval and pick a reviewer, and approvers can approve the request, send it back for more edits or withdraw it. In the free version, the permission levels are your approval step: give Publish only to the people who should sign off.

Every publish records a version, and the Versions screen shows who published it and when, as covered in the Tag Manager version history article. Still, none of these histories stores the reason or the expected result, and each lives in a different tool. That is the gap the shared log fills.

How do I isolate which change caused a drop when several changes overlap?

Put every change from the days before the drop on one timeline, rule out tracking first, then test one hypothesis at a time. A change process makes this fast because the record already exists.

1. Date the shift. Find the first day the metric moved, in daily data, not weekly.
2. Pull every change from the 7 to 14 days before. Your team's edits, client edits, platform automations, site releases, tracking changes and external events.
3. Rule out tracking. Check whether a tag, container version, conversion action or form changed. If clicks and sessions are steady but conversions fall, tracking is the first suspect.
4. Test one hypothesis at a time. Roll back or compare one change, wait for enough data, then move to the next.

Example: on Thursday a client asks why form leads fell 35% on Wednesday. The team opens the log. Monday: target CPA lowered from €45 to €40, approved and logged. Tuesday: the client's developer shipped a redesigned contact form, and a new GTM version went live the same day. Clicks were flat, so tracking came first: the form's ID had changed and the lead tag no longer fired. The team fixed the trigger, leads came back on Friday, and the CPA change was judged a week later on clean data.

With every change on one timeline, the form release and the GTM publish sit right before the drop, next to the bidding change, so you check the likely cause first instead of guessing. In Changeline, Google Ads edits can be imported automatically and the team logs site, tracking and CRM changes on the same project timeline.

For the write-up after an incident like this, use the method in the marketing post mortem article.

How do I review changes without turning the process into bureaucracy?

Hold a 20-minute weekly review of the logged changes and cut any step nobody uses. A process that slows every edit gets skipped, and a skipped process is worse than none, because people then trust an incomplete log.

• Review by result, not by person. Go through the week's high and medium-risk changes and mark each one: worked, didn't work, too early. The log becomes a learning record instead of a blame list.
• Let low-risk changes flow. If approvals pile up, the tiers are too strict. Move routine edits down a tier.
• Spread changes out. When two high-risk changes land on the same campaign in the same week, delay one. Overlap is what makes results impossible to explain.
• Use the record in client reporting. A monthly report that lists what changed and why answers most what-happened questions before anyone asks them.

On Changeline's paid plans, the Report button builds a PDF of the changes logged or imported in a chosen period, which fits the monthly review; the plans page lists what each plan includes.