Google Ads import into Microsoft, de-duplicated

The import button takes five minutes and creates three months of quiet waste. In the accounts we have inherited after somebody used it, between a fifth and a third of imported keywords ended up competing with keywords already live elsewhere in the same Microsoft account, negatives written for Google's close variant behaviour were over-blocking Microsoft traffic, and conversion goals pointed at a UET tag that was firing on two pages out of forty. None of that shows up as an error. It shows up as a channel that never quite worked. What we inventory before anything moves: A full export of the destination account as it stands: keywords by match type and ad group, negative lists and their assignments, account-level negatives, shared budgets, ad schedules, location targeting method, and which conversion goals are currently flagged to count in the Conversions column. Then the same from Google. The two lists are compared before a single entity crosses over, because the decision that matters is which campaign should own a term, and that decision needs conversion history from both sides. De-duplication: Ownership is assigned by evidence. Where the same term exists in two campaigns, the one with real conversion history keeps it and the other gets it as a negative at the right match type, rather than a bulk delete that throws away quality history. Negative lists are rebuilt rather than imported: Google negatives are written against Google's close variant behaviour, and dropping them wholesale into Microsoft blocks traffic that would have converted. Duplicate negatives across an account and a list are the single most common cause of a Microsoft campaign that inexplicably stops serving. Conversions and attribution: We rebuild goals natively rather than importing them. The UET tag is verified across every page that matters, revenue is passed with the goal so value-based bidding has something to read, and the attribution window and model are stated explicitly rather than inherited. Google and Microsoft do not count the same conversion the same way, and the reconciliation belongs in your reporting document on day one, not in an argument in month three. What we change on the way in: Daily budgets do not come across at Google's numbers. Microsoft's query volume in Swedish is a fraction of Google's, and budgets set at parity spend on the wrong half of the account. We model them against achievable impression share instead. Location targeting is set to physical presence unless there is an argument for intent, because the imported default reaches people merely reading about the location. Performance Max campaigns we rebuild rather than import; the import carries asset groups without the signals that made them work. Scheduled imports: We do not leave one running daily. A standing auto-import quietly reintroduces entities you paused, overwrites bids you set deliberately in Microsoft, and makes the account a mirror rather than a channel. We set a monthly manual import with a written exclusion list, and the first thing we do on any inherited account is switch the automatic one off. Not included: Fixing the Google account. If the source is a mess, the import inherits the mess and we will tell you what needs doing there, but repairing it is separate work. Also excluded: Microsoft Shopping feeds, audience network setup, and ongoing management after handover. Not for: Accounts likely to spend under roughly 5,000 EUR a month on Microsoft, where the honest advice is a small manual build rather than a migration. And not for anyone who wants Microsoft to be a mirror of Google, which is a preference we will argue with rather than execute.