Engines and sync
An engine account is one account on one search engine, linked under a client account in SA360. Linking means SA360 can read the account's campaigns, ad groups, keywords, ads and settings, report on them, and send edits back. The relationship is two-way but not symmetrical: SA360 is the management layer, the engine is the system of record for what actually serves, and sync is the process that keeps the two aligned. The tutorial on moving paid search into Search Ads 360 puts linking and the first sync in order with the Floodlight check, labels and the bidding handover.
| Engine | How it is managed | What sync does | What stays in the engine |
|---|---|---|---|
| Google Ads | Near-native: the current experience shares infrastructure with Google Ads, so most campaign types, settings and assets are available | Changes made in either place are reflected in the other; the help centre describes the cases that still need a sync | The small set of features the help centre lists as unsupported |
| Microsoft Advertising | Through the engine's interface: a documented subset of campaign types, settings and ad formats | Scheduled and on-demand sync pulls engine changes in; SA360 edits are sent to the engine and can be rejected by it | Settings outside the supported subset, including some audience and automation options |
| Yahoo! Japan Ads and other listed engines | As for Microsoft: a documented subset per engine | Scheduled and on-demand sync | Everything not on that engine's supported list |
Three documented behaviours shape day-to-day work. First, an edit made in the engine does not exist in SA360 until sync reads it, so a report run before sync shows the old state. Second, an edit made in SA360 to a non-Google engine is sent to that engine and can be rejected, for policy, format or limit reasons, in which case SA360 records the error rather than silently dropping the change. Third, sync is not instant and Google does not publish a guaranteed delay; the help centre describes the cadence and the manual sync option, and the practical rule is to sync before any bulk work and before any report that will be acted on.
The previous version of SA360 ran its own trafficking and sync pipeline for every engine, Google Ads included; the current experience removes most of that step for Google Ads while keeping it for the others. Anyone moving from the old product should expect Google Ads edits to appear faster and the other engines to behave much as before.
Managing campaigns across engines
Campaigns are built in SA360 the way they are built in Google Ads: a campaign with settings and budget, ad groups with keywords and ads, assets attached at the appropriate level. The difference is the engine account chosen at creation, which decides which engine the campaign serves on and which features are available in the editor. The structure principles do not change (group by intent, consolidate where automated bidding needs data, split only where control is genuinely needed), and the Google Ads account structure guide applies to each engine account inside SA360.
Mirroring a campaign from one engine to another is where most teams expect a button and find a process. The reliable path is a bulk download of the source campaign, an edit of the engine account column and anything engine-specific, and an upload into the target account. Three things need attention in the edit.
- Match types. Both engines use broad, phrase and exact match, but their close-variant behaviour and their treatment of broad match are not identical, so a keyword set tuned for one engine's matching is not automatically right for the other. The keyword match types guide explains the Google side; Microsoft's own documentation covers its matching.
- Ad formats and assets. Responsive search ads exist on both engines, but the supported asset types and limits differ, and an upload that exceeds the target engine's limits is reported as an error rather than quietly trimmed.
- URLs and tracking. Landing pages, tracking templates and custom parameters must produce the click identifier each engine uses, or Floodlight conversions will not join to the clicks. This is the single most common cause of a mirrored campaign that appears to convert on one engine and not the other.
Budgets stay at the engine's campaign level and are set per engine; SA360 does not pool a budget across engines on its own. The help centre documents budget management features that pace spend across a group of campaigns, and whether they fit a given account depends on its engines and bidding; treat them as a separate decision from mirroring.
Negative keywords, ad schedules, location targets and device settings are set per campaign and are part of the mirror. Shared negative lists are engine-level objects: a list created in Google Ads is not applied to Microsoft campaigns, so the same exclusions have to exist on each engine, which is one of the jobs that labels and bulk work make tolerable.
Labels, templates and bulk work
Labels are the glue of cross-engine management. A label is attached to campaigns, ad groups, keywords or ads on any engine, and then used to filter reports, to select objects for a bulk edit, and to pick campaigns for a bid strategy. A campaign on Google Ads and its mirror on Microsoft Advertising with the same label (a product line, a market, a funnel stage) can be reported on as one row and edited as one set. Labels complement rather than replace names: the fields a name carries, in the order the naming conventions guide describes, should be identical on every engine so that the name alone sorts and filters consistently, and labels carry the dimensions that change more often than names should.
Templates in the current experience come in two kinds. Tracking templates and custom parameters, set at account, campaign or ad group level, define how final URLs are decorated, and they are the mechanism that keeps click identifiers and campaign parameters consistent across engines. Inventory templates, covered in the next section, define how feed attributes become campaign objects. Both are worth treating as governed configuration: a template edited casually changes every URL or every generated ad beneath it.
Bulk work has three documented routes. Bulksheets (download the objects you want as a spreadsheet, edit, upload) handle large structural changes and are the route for cross-engine mirroring. In-interface bulk edits handle a filtered set of objects at once, such as pausing every keyword with a label or changing a setting across campaigns. Automated rules run a condition and an action on a schedule, such as pausing ads whose approval status changes or flagging a campaign that is limited by budget. Every route writes to change history, which is the audit trail when something moves and nobody admits to it.
The cost of bulk work on a layer is that mistakes are bulk too. A download that is edited and uploaded hours later can overwrite changes that sync brought in between; a filter that is slightly too wide pauses a market. The working rules are to sync immediately before downloading, to keep the edited columns to a minimum, and to upload to one engine account first and check it before uploading to the rest.
Inventory-driven campaigns
Inventory management generates search campaigns from a product feed and keeps them current as the feed changes. A set of inventory templates defines how feed attributes (product name, brand, category, price, availability) become campaigns, ad groups, keywords, ads and sitelinks; SA360 reads the feed, creates the objects the templates describe, updates ad text and keywords as attributes change, and pauses items that fall out of stock. The result is a search presence that follows the catalogue without anyone editing ads by hand.
In the current experience the documented feed source is Google Merchant Center, so the product data has to meet Merchant Center's requirements before inventory management can use it; the previous version had its own feed set-up, and teams migrating should not assume old feeds carry over. The help centre also lists which engines inventory campaigns can be generated for and which object types each template kind can produce; the list is narrower for non-Google engines, as it is for everything else.
Inventory campaigns suit catalogues with many items, frequent price or stock changes, and a long tailA specific, usually longer query searched rarely on its own but often as a group. of product queries where hand-built keywords would never keep up. They suit less well a small range with stable messaging, where a template adds a layer of indirection for little benefit, and brand or category campaigns where the ad copy is a creative decision rather than a feed attribute. The usual design is to keep inventory campaigns for the long tail and hand-built campaigns for the head terms, labelled so that reports can separate the two.
Two failure modes recur. A feed that stops refreshing leaves ads live for products that no longer exist, which is the first thing to check when an inventory campaign's conversion rate collapses. And a template rule that matches more feed rows than intended generates campaigns nobody planned, with budgets attached; a template change should be previewed and the generated objects counted before it goes live.
Mirroring a product-category structure onto Microsoft Advertising
A sports retailer runs 48 non-brand Google Ads campaigns, one per product category, and wants the same structure on Microsoft Advertising, where it currently has 6 catch-all campaigns. The team syncs both engine accounts, downloads the Google campaigns as a bulksheet, and in the sheet changes the engine account, removes the 11 Google-only asset types and audience segments, and reviews the 3,900 keywords against Microsoft's matching behaviour, dropping 240 broad-match terms that depend on Google's audience signals. Tracking templates are checked so that every final URL carries the Microsoft click identifier. The upload goes to one engine account first; 14 ads are rejected for exceeding Microsoft's headline count and are trimmed. Each campaign pair carries the same label, so a single cross-engine report shows category performance on both engines from the first week, and the 48 Microsoft campaigns join the existing target CPA portfolios by label.
Team workflow and permissions
Access in SA360 is granted at manager, sub-manager or client account level and inherited downwards, with roles that range from administrative control to read-only; the help centre lists what each role can do. The design question is where to grant: a regional lead needs the sub-manager account, an engine specialist needs the client account or a single engine account, a finance analyst needs read-only access at the top. Granting more widely than that is how a bulk upload for one market ends up in another.
The workflow that keeps the layer honest has four parts. An agreed answer to "which objects are edited where": in most teams, SA360 for anything that spans engines or feeds a bid strategy, the engine for engine-only settings, and never both for the same object. A sync habit before any bulk work and before any report that will be acted on. A review of change history on a cadence, so that drift between the engines and SA360 is caught by looking rather than by a performance surprise. And a KPIA metric tied directly to an objective. set that is defined once in the cross-engine report, with the same cost per acquisition (CPACost of each conversion.) basis on every engine, so that the team is not comparing a Floodlight figure on one engine with an engine-tag figure on the other.
Notifications and automated rules can carry some of this: a rule that flags a campaign with no impressions for a day, a notification when an upload is rejected. They are worth setting up early, because the alternative is discovering a failed upload at the end of the month, after the report has gone out.
Common mistakes
- Treating SA360 as the system of record. The engine is; SA360 is a layer that reads and writes to it. Reports run before sync, and uploads made without sync, describe a state that no longer exists.
- Mirroring campaigns without adjusting for the engine. Match types, asset limits and audience features differ, and an upload that ignores them is rejected or performs oddly.
- Forgetting tracking on the second engine. Click identifiers that work on Google Ads do nothing on Microsoft unless the tracking template produces that engine's parameter; conversions then appear to vanish on one engine.
- Labels applied on one engine only. A label that exists on the Google campaigns and not on their Microsoft mirrors makes cross-engine reports lopsided.
- Bulk edits from stale downloads. A sheet downloaded in the morning and uploaded in the evening overwrites whatever sync brought in between.
- Inventory templates nobody owns. A template change regenerates every object beneath it; without an owner and a preview habit, campaigns appear with budgets attached and nobody knows why.
- Over-wide access. Users granted at the manager level who only needed one client account, and a bulk upload that lands in the wrong market.
The SA360 hub sets this guide against bidding and measurement; the bid strategies guide is the next step once the structure is in place.
