Handing ad operations to an outside team works when you define the work, write the handoff down and test it on a small set of campaigns first. It fails when a loosely described media plan is thrown over the fence. This guide is a step-by-step how-to for agencies that have already decided to outsource part of their ad ops and now need to set it up well.
If you are still deciding whether to outsource at all, start with in-house vs outsourced ad operations. If you want to know what a provider like us does, see our outsourced ad operations service page. This post focuses on the how.
1Step 1: Decide whether the work is ready to hand off
The work you outsource should be repeatable and documentable. A useful test: could a competent stranger do it from a written checklist and access to your accounts? If yes, it is a good candidate. If the work depends on judgment calls that live in one person's head, document it first or keep it in-house.
| Work type | Usually a good fit to hand off | Usually keep closer to your team |
|---|---|---|
| Trafficking and setup | Ad server and DSP campaign setup, tag and creative trafficking, naming and tracking conventions | Final approval of unusual client requests |
| QA | Pre-launch checks, click-through and pixel testing, device and placement checks | Sign-off on brand or legal-sensitive creative |
| Pacing and optimization | Daily pacing checks, budget adjustments within agreed rules, flagging underdelivery | Channel strategy, KPI trade-offs, client conversations |
| Reporting | Recurring reports, data pulls and reconciliation between platforms | Insights and recommendations that need client context |
| Billing and reconciliation | Invoice checks against platform spend | Client billing terms and disputes |
Outsourcing is usually a weak fit for an agency with very few campaigns, no written processes, or work that changes completely every week. In those cases, tighten your own process first.
2Step 2: Write the scope as a task list, not a service name
"Programmatic support" means different things to different people. Replace it with a list of tasks and owners. For each task, record what triggers it, who does it, what done looks like and how long it should take.
A simple scope matrix can be a spreadsheet with one row per task and four columns: task, owner (your team or the provider), trigger (for example "signed IO received") and output (for example "line items built, QA checklist completed in the ticket"). Pay attention to the edges: who requests creative changes, who talks to the client when something goes wrong, and who owns the platform accounts.
Be specific about the platforms. A team that is strong in Google Display & Video 360 may not know your Campaign Manager 360 setup, The Trade Desk or Amazon DSP. List them, along with the access level each person needs.
3Step 3: Choose an engagement model
Providers generally price in one of a few ways. Which suits you depends on how predictable your volume is. Our pricing models guide goes into the trade-offs in more detail.
| Model | How it works | Best when | Watch out for |
|---|---|---|---|
| Dedicated resource | A named person or small team works your queue | Steady, predictable volume | Paying for idle time in slow weeks |
| Per-task or per-campaign | A set fee per campaign setup or task | Volume varies, tasks are standardized | Unclear definition of what counts as one task |
| Percent of media spend | A fee tied to managed spend | Provider also manages optimization | Cost rises with spend even if the work does not |
| Retainer with a work allowance | Fixed monthly fee covering an agreed amount of work | You want a predictable bill | What happens when you go over the allowance |
Whichever you choose, ask what is included, what is billed extra (rush work, weekends, extra reporting) and what notice is needed to scale up or stop.
4Step 4: Agree measurable service levels
A service level agreement (SLA) turns "fast and accurate" into numbers you can check. The most useful ones are usually turnaround time per task type, a first-time-right rate on QA, reply time on questions and the time to escalate a delivery problem. Set the numbers with your provider based on your workflow instead of copying a template. Our ad operations SLA guide lists the metrics and how to track them.
Two points that are often missed. First, define the clock: when does a request count as received, and what counts as complete? Second, decide what happens when the SLA is missed. A review meeting and a root-cause note are more useful than a penalty clause alone.
5Step 5: Set up access and security
- Use named logins in your ad server, DSPs and analytics tools instead of shared credentials, so actions are traceable and access can be removed.
- Give the least access needed. A trafficker may not need billing permissions, for example.
- Decide how credentials, API keys and client data are shared, and where files are stored.
- Put confidentiality and data-processing terms in the contract. If client data is covered by GDPR or similar rules, make sure the contract and your client agreements allow it.
- If you want white-label work, agree on email domains, ticket naming and how the provider introduces itself to clients.
6Step 6: Write the handoff package
This is the document that decides most of the outcome. A good handoff package has the following parts.
- Intake form: a standard form for every new campaign with client, objective, flight dates, budget, platforms, targeting, creative list, tracking requirements and approvers.
- Naming and tagging conventions: campaign, line item and creative names, UTM rules and how they map to reporting.
- QA checklist: the checks to complete before launch, with a place to record results. See our ad trafficking QA checklist for a starting list.
- Escalation path: who to contact for what, and how quickly.
- Client notes: brand rules, past issues, preferred contacts and anything that surprised you before.
Writing this will expose gaps in your own process. That is useful, and usually worth doing even if you later decide not to outsource.
7Step 7: Run a pilot before moving everything
Start with a small number of stable, lower-risk campaigns for the first 30 to 60 days. Avoid your most complex client launch or a campaign with a hard date. Agree before you start what you will measure, for example turnaround time, errors found in QA, errors found after launch, number of clarifying questions and how much of your own team's time it takes to supervise.
Review weekly during the pilot. Look at every error and ask whether it came from the brief, the process or the execution, and fix the cause in the documentation. Our guide to onboarding an outsourced ad ops team has a week-by-week outline.
8An illustrative example
For example, an agency running programmatic display for several mid-size clients might hand off three things in the pilot: campaign setup in DV360, ad server trafficking in Campaign Manager 360 and a weekly pacing report. The agency keeps client communication and optimization strategy. After a month, the agency reviews a small sample of the tickets: was the intake form complete, did QA catch the issues, were the reports on time? If the intake form is the main source of delays, the next step is to fix the form, not to change providers. This example is illustrative and not a result from a specific client.
9Common mistakes
- Moving everything at once instead of piloting.
- Giving vague briefs and expecting the provider to fill gaps in the same way your team would.
- Not having a named person on your side who owns the relationship and reviews the work.
- Measuring cost only. Check quality and the time your own team spends on rework.
- Skipping the exit plan. Keep documentation and account ownership on your side so you can change provider or bring work back in-house.
10Next step
If you are ready to scope a pilot, our outsourced ad operations service page explains how we work, and we also cover agency-specific setups in our white-label ad operations guide. Contact us with your platforms and campaign volume and we will suggest a pilot scope.
