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

Standard Operating Procedures

Compromised Marketing Account Response SOP

A standard procedure for a social, ad or email account in the wrong hands: confirm, pause, recover through official routes, remove intruders and log it.
SOPintermediate11 min readUpdated

Owner, trigger and scope

The account owner runs the response: the person the access register names as accountable for the affected account, or their deputy admin. The social media lead decides public statements under the crisis plan. The IT or security contact handles devices, email accounts, password resets and anything that suggests the problem reaches beyond one marketing account. The data protection owner decides what follows if personal data may have been exposed.

The trigger is any sign that a social, advertising, email or other marketing account is being used by someone who should not: posts or ads nobody on the team made, an unknown admin, sign-in alerts, changed recovery details or payment changes. The tutorial on securing social media accounts calls an account taken over by someone else a crisis in its own right and hands it to the crisis plan; this procedure is the account side of that crisis, for every kind of marketing account, while the social media crisis plan runs the public side.

Neighbouring situations have their own pages:

SituationWhere it is handled
A wrong post, ad or page the team itself publishedLive ad and page error response SOP
Granting, changing or removing access in the normal course of workMarketing tool access requests SOP
A business profile the business cannot manage because it lost track of who holds itLocked out of a business profile
A crisis affecting the whole business, not one accountLeading marketing through a business crisis

The signs often first show up in the social inbox, as followers asking about posts the team never made; the social media inbox triage SOP escalates them here. The SOP library routes the other incident procedures. The academy checkpoint on hijacked accounts, live errors and incident programmes tests confirming and recording a takeover, the official recovery routes and the checks before restarting.

Before you start: inputs and access

InputOwnerNotes
The access register with owners, admins and recovery details for each accountAccount ownerWho should have access, at what role, and which email and phone recover it
The business vault holding credentialsIT or security contactCredentials for accounts that do not support individual users
Each platform's official recovery and reporting routeAccount ownerReached from the platform's help centre directly, never from a link in a message
The crisis plan with holding statementsSocial media leadSeverity levels, who speaks and the approved wording
Payment and billing contacts for each ad accountFinanceWho can check charges and contact the card issuer or bank

Most of these come from the security tutorial's account register and the guide to reviewing access to marketing tools. If they do not exist when an incident starts, the response still runs, but slower: the gaps are the first finding for the review. The marketing incident contact sheet template keeps the account owner, the deputy admin, the recovery and billing routes and the people to tell on one page that can be read without logging in to anything.

The procedure

  1. Confirm the signs and record what you see

    Check that the activity is genuinely not the team's: ask the people with access whether anyone posted, created the ad or changed the settings, and check the agency. Then record what you see before anything is changed, with screenshots and times: the posts, ads or messages, the unknown users, the alerts, the changed recovery details. The security tutorial lists what each account should look like, with its owner, admins and recovery details, so the difference is visible. Open the incident log now, and make its first line the time and date you became aware of the signs; every later step adds a timed line to it. That first line is evidence of when the signs appeared, not the start of any regulator-notification deadline. If a personal data breach turns out to be reportable, the 72-hour deadline described in step 2 counts from when the business becomes aware of that breach, and the data protection owner decides when that was.

  2. Pause the publishing queue and stop ads you can still reach

    Pause scheduled posts in the scheduling tool and stop any campaign the team can still reach, starting with anything the intruder created and anything spending. The crisis plan tutorial includes pausing the queue among the first-hour actions and plans how it restarts. If the intruder has used the account to promote something harmful or misleading, the brand's Brand Safety is at stake as well as its money, which is a reason to move quickly here even before access is fully recovered.

    If audiences, email lists, lead forms or customer messages can be reached from the account, tell the data protection owner now rather than waiting for step 7, giving them the time you noted in step 1. A personal data breach that is reportable under UK or EU GDPR has a short regulator-notification deadline, 72 hours, counted from the moment the business becomes aware of the breach, so the clock may already be running. The data protection owner decides whether the incident is reportable and what to tell the regulator; see the ICO's guidance on reporting a personal data breach or the equivalent regulator in the business's own country. This procedure is not legal advice.

  3. Recover the account through the platform's official route

    Use the recovery and reporting route each platform documents in its own help centre: the Meta Business Help Centre, Google Ads Help, the TikTok Ads Help Center and the equivalent for email and other platforms. The screens, forms and outcomes live there and change, so this procedure does not quote them. Where the business owns its assets through a business-level account, as the business portfolio tutorial and the Business Center setup tutorial recommend, a second admin who still has access may be able to act first. The portfolio tutorial is explicit about why that matters: an account the business does not own cannot be secured by any later step.

  4. Change credentials, sign out other sessions and restore two-factor and recovery details

    Once in, change the password of the affected login and of any shared credential it used, from a device the IT or security contact trusts, and sign out every other session where the platform allows it. Switch two-factor sign-in back on, or on for the first time, and move the recovery email and phone to business-controlled addresses, as the security tutorial sets out. If the way in was a personal email account or device, the IT or security contact secures that too, since the same route may reach other accounts. For an email platform or any account connected to a CRM or other tools, also check for forwarding rules and automatic rules the intruder may have set up, which can keep copying mail or leads out after the password is changed, and revoke API tokens and app passwords issued from the account, reissuing only the ones the business still needs.

  5. Remove unknown users and check the admins against the register

    Compare every user, partner and connected app on the account with the access register and remove anything that does not belong, including anything the intruder added. The guide to reviewing access to marketing tools shows how to match each account to a current person, supplier or integration and record each removal. Make sure at least two named admins remain, both employed by the business. If a former agency or freelancer still held access, the agency offboarding checklist covers what should have been removed and rotated when the relationship ended. New access requested during recovery goes through the marketing tool access requests SOP, not around it.

  6. Check ads, spend, payment methods and anything sent in the brand's name

    Go through every campaign, ad, post and message created or changed while the account was out of the team's hands, and stop or remove what the intruder made. Check spend and the payment methods on each ad account; the guide to manager accounts and access treats billing details and payment profiles as belonging to the business, with finance owning them, and that is who checks the charges. Report any unauthorised spend through the platform's own route and tell the card issuer or bank where finance decides to. Check the email platform and direct messages for anything sent to customers in the brand's name.

  7. Tell customers and the data protection owner where needed

    If the intruder posted, advertised or messaged in the brand's name, the social media lead decides the public statement under the crisis plan, usually a short Holding Statement that says the account was misused and what people should ignore. If customer lists, messages, form leads or other personal data may have been seen or taken, the data protection owner should already have been told at step 2; if these checks are the first sign of it, tell them straight away. Either way, pass them what the checks in steps 4 to 6 found. They decide what follows under the business's own obligations, which may include the GDPR. The marketing team does not decide those questions, and this procedure is not legal advice.

  8. Restart in order and log the incident for review

    Restart publishing and campaigns in a deliberate order, checking each scheduled post and ad before it goes back live; the business crisis guide sets out why everything scheduled is reviewed before anything restarts. Then complete the incident log and decide on a review. The tutorial on reviewing a marketing error builds a timeline from records and asks why until the cause is a process, using the five whys framework, so the finding is a missing control rather than a person to blame.

A skate shop's hijacked page (illustrative)

Illustrative example — not a real result

Giveaway posts, an unknown admin and an ad set stopped in time

Rimvoskel Skate Shop, an invented skateboard and clothing shop, sees 3 giveaway posts on its page on a Saturday morning that nobody on the team made, and a follower asks in the comments whether the prize is real. The account owner checks with the 2 staff and the freelancer who post, confirms none of them made the posts, and screenshots the posts and the portfolio's user list, which shows an admin nobody recognises. The scheduler is paused. The shop manager, the second admin, still has access and starts recovery through the platform's documented route from the help centre, then removes the unknown admin and the posts. Passwords are changed, two-factor sign-in is switched back on for the freelancer, who had turned it off, and recovery details are moved to the shop's own inbox. The check of the ad account finds a new ad set promoting the fake giveaway with a daily budget of £150, created 40 minutes earlier; it is stopped before the first charge clears and reported through the platform's route. The social media lead posts a holding statement saying the giveaway was not the shop's and asking followers not to share their details. No customer data was held in the account, which the data protection owner confirms. The incident log records 11 timed actions, and the review finds the cause in a password reused by the freelancer; the action goes into the next access review.

QA checks

Before the incident is closed, the account owner confirms each check below. If one fails, the incident stays open.

  • Only people on the access register hold access, with at least two named admins.
  • Recovery details and two-factor sign-in point to business-controlled addresses and devices.
  • No ads, posts or messages created by the intruder remain live on any account they reached.
  • Billing has been checked and any unauthorised spend reported through the platform's route.
  • The incident is in the log with a decision on review and, where there is one, its owner.

Hand-offs and records

The procedure hands on four things. The social media lead takes the public statement and any follow-up under the crisis plan, and the community team receives the wording for replies. Finance receives the spend and payment check and any report to the card issuer. The data protection owner receives everything known about personal data that may have been exposed and decides what follows. Marketing operations receives the review actions, so the next access review and the marketing tool access requests SOP close the gap the intruder used.

The lasting record is an incident log entry: when the signs were seen and by whom, what the intruder did, each containment and recovery action with its time, users removed, spend and messages checked, statements made and the review decision. Keep the screenshots with it. The log is what the review reads, and what the next person looks at if the same account is targeted again. The academy exercise on handling a hijacked account and a wrong ad price for a climbing centre practises running this response beside a live ad error found on the same weekend, with both incidents logged for one review.

Common mistakes

Each of these turns a contained incident into a longer one.

  • Changing things before recording them. Once the intruder's posts and users are removed, the evidence of what happened goes with them.
  • Using a recovery link from a message. Reach the platform's help centre directly; a message asking for passwords or codes may be how the account was taken.
  • Recovering one account and stopping. The same password, email or device may reach the other accounts; check every account on the register that shares them.
  • Forgetting the ad accounts. An intruder with access to a page may also reach the ads and the payment method.
  • Restarting the queue unchecked. Scheduled posts written before the incident may read badly after it.
  • Deciding personal data questions in the marketing team. They belong to the data protection owner.
  • Closing without a review. The gap that let the intruder in is still open until someone owns the fix.

Frequently asked questions

Should we pay or reply to someone who offers to return the account?

No. Recovery runs through the platform's own documented route, and anyone offering a shortcut, whether the intruder or a stranger claiming to help, is more likely to make things worse. Messages that claim to come from the platform and ask for passwords or codes may be how the account was taken in the first place. Use only the help centre's official recovery and reporting routes, reached directly rather than through a link someone sent.

Do we need to tell our followers or customers?

Often, yes, when the intruder posted, advertised or messaged in the brand's name, because people may have acted on it. The social media lead decides the public statement under the crisis plan, usually a short holding statement that says the account was misused, that the posts were not the brand's and what people should ignore. Whether anything further is needed because personal data was exposed is a question for the data protection owner, not for the marketing team.

How do we stop it happening again?

Review the incident and act on what it finds. Most takeovers trace back to a gap the security tutorial already names: a single admin, a shared password, recovery details on a personal phone, two-factor sign-in left off, or access never removed from someone who left. The review turns the gap into an action with an owner, and the next access review checks it was done.

Terms used on this page

Frameworks

Where this page is used

Part of these learning paths

Also related

Sources

  1. Meta Business Help Centre — Meta Business Help Centre
  2. Google Ads Help — Google Ads Help
  3. TikTok Ads Help Center — TikTok for Business

Related