Buying on one retail media network is a channel. Buying on five is an operations problem. Each network you add brings its own console, ad formats, creative rules and reports, and the team that was comfortable running Amazon Ads suddenly finds itself logging into Walmart Connect, Instacart and a regional network dashboard every morning, rebuilding the same spreadsheet to answer the same question: where is the money working?
In short: managing several retail media networks well means accepting that they will never be identical, then building a layer above them. That layer is a shared taxonomy and naming convention, a budget allocation method based on goals rather than habit, reporting normalized to common definitions, a single source of truth for creative specs, a clear staffing model, and a regular governance cadence that reviews every network against the same plan.
1Why Every Network Is Different
Retail media networks grew up separately, so they share concepts but rarely share details. The differences that cause the most work are:
| Dimension | What varies |
|---|---|
| Platforms | Some networks run their own self-serve console, some license technology from a vendor, and some are still largely managed-service with insertion orders and email. |
| Formats | Sponsored products exist almost everywhere, but sponsored brand units, video, display placements and brand stores vary in name, layout and availability. |
| Data rules | What audiences you can build, whether you can export data, and whether clean-room access is offered all differ by network. |
| Creative specs | Image sizes, file weights, headline lengths, logo rules and review times change from one retailer to the next. |
| Attribution windows | One network may credit a sale up to a set number of days after a click, another after a view, and some count sales of any product from the brand rather than only the advertised one. |
| Reporting | Metric names, currencies, time zones and levels of detail rarely match. |
None of this is going to be standardized for you. The practical answer is to standardize your own side.
2Build a Common Taxonomy and Naming Convention
A taxonomy is the set of labels you apply to every campaign, whatever network it runs on. It is what lets you later ask "how much did we spend on branded search for this product line across all networks?" and get an answer without manual tagging. Define it once, in a shared document, before the next campaign goes live.
A workable campaign name usually carries, in a fixed order:
- Network and country, for example AMZ-IN or WMT-US.
- Brand and product line, using the same short codes everywhere.
- Ad type, such as sponsored products, sponsored brands, display or offsite.
- Objective or funnel stage, such as defend, conquest or new-to-brand.
- Targeting type, such as branded keywords, category keywords, competitor or audience.
- Start date or flight, so old and new versions don't get confused.
Keep the codes short, use one separator, and never reuse a code for two meanings. Where a network limits name length, the order of fields tells you which ones to drop first.
3Allocate Budget Across Networks on Purpose
Without a deliberate method, budget tends to follow the network that is easiest to spend on or the account manager who asks most often. A better starting point is your own sales: how much of your business each retailer represents, and where you want that share to grow. From there, set a role for each network. One may be about defending share on your largest retailer, another about launching a range on a newer marketplace, and a third about reaching new customers through offsite audiences.
Split each network's budget into an always-on base, mainly sponsored products, and a flexible layer for display, offsite and seasonal pushes. Review the flexible layer monthly and move it toward the networks beating their goals, after checking that stock and pricing were not the reason others fell short. Our network selection and budget planning service builds this plan across every network you buy.
4Normalize Reporting Before You Compare
Putting ROAS from three networks side by side looks like analysis, but if each network uses a different attribution window and a different definition of an attributed sale, it is comparing different things. Normalizing means agreeing your own definitions and mapping every network's data to them.
- Pick standard metric definitions: spend, attributed sales, ROAS, new-to-brand sales and share of voice, with a note of how each network's version maps to yours.
- Record each network's attribution window and click versus view rules next to the data, so nobody compares them blindly.
- Align currency, time zone and week definitions before combining figures.
- Apply your taxonomy so every row can be grouped by brand, ad type and objective.
- Automate the pull through APIs or connectors where networks offer them, rather than downloading files each week.
A single dashboard built on those rules replaces the Monday spreadsheet. Our reporting and analytics team builds these cross-network views, and when you need to separate attributed sales from incremental ones, see our article on clean rooms, incrementality and attribution.
5Manage Creative Specs in One Place
Creative rejections are one of the most common reasons a multi-network campaign launches late on some networks and on time on others. Keep a single spec sheet that lists, for each network and format, the dimensions, file types and weights, copy limits, logo and claim rules, and the typical review time. Update it whenever a network changes a requirement, and make it the first thing a designer opens.
Then design for adaptation rather than one-off builds. A master set of assets, with layered source files and agreed safe zones, can be resized and re-cut for each network far faster than starting again each time. Build in a QA check against the spec sheet before anything is uploaded. This is the sort of volume work our creative operations team takes on, and it matters most for Sponsored Brands and onsite display ads, where formats vary the most.
6Choose a Staffing and Operating Model
There are two common ways to organize the team.
| Model | How it works | Strength | Trade-off |
|---|---|---|---|
| Network-specialist | Each person owns one or two networks and knows them deeply. | Faster on their own network. | Creates silos. |
| Function | People own a function, such as search management, display trafficking, creative or reporting, across all networks. | Keeps things consistent. | Needs broader training. |
Many teams use a mix: a function model for reporting and creative, and network specialists for day-to-day search management.
Whatever the model, document each network's quirks in a shared runbook so knowledge doesn't leave with one person, and use named logins rather than shared ones. If you lack the people to cover every network properly, outsourcing parts of the work is an option; our article on in-house versus outsourced ad operations sets out the trade-offs. Retailers facing the same problem from the other side, running their own network for many advertisers, can look at our ad operations for retail media networks.
7Set a Governance Cadence
A cadence turns all of the above into a habit. A simple rhythm that works for most brands and agencies:
| Cadence | Focus |
|---|---|
| Daily | Pacing, stock alerts and any disapproved ads, on every network. |
| Weekly | Search term and bid reviews, budget shifts within each network, and a short cross-network performance check. |
| Monthly | Normalized reporting against goals, reallocation of the flexible budget, and spec sheet and taxonomy updates. |
| Quarterly | A review of each network's role, whether to add or drop a network, and the next quarter's plan. |
Running several networks never becomes simple, but it can become predictable. To see how we run multi-network programs end to end, visit our retail media marketing page, or start with a free retail media audit of your current networks.
