Microsoft Ads import from Google: keep the structure or rebuild from scratch?

Microsoft Ads can import a Google Ads account in one go, and people still argue about whether that's a shortcut or a trap. My view: import the structure, not the settings. Campaigns, ad groups, keywords and ads carry over fine, but the parts tuned for Google often don't behave the same way. Smart Bidding targets built on Google's conversion history need checking against Microsoft's own UET conversions, because a goal imported without enough volume can sit in learning for weeks. Broad match also reaches looser queries on Bing, so a keyword list that was tight on Google can spend on traffic you never meant to buy. Say a dental clinic imports 400 keywords and gets 12 tracked leads a month. The bidding strategy is the first thing I'd change, not the keyword list. Does anyone import everything and fix it afterwards, or do you rebuild from scratch?

Agreed on bidding, but with 12 leads a month I'd skip Smart Bidding for now. Target CPA and Maximize conversions need a steady flow of conversions to learn from, and 12 a month spread across 400 keywords gives the algorithm almost nothing to work with. It sits in learning and keeps spending while it waits. Manual CPC or Maximize clicks with a sensible daily cap keeps you in control until volume is high enough to trust a target.

For local services, a good rule of thumb is to import the structure but hold off on a Smart Bidding goal until Microsoft's own tag has recorded around 30 conversions in a month. Below that, treat the imported account as a test and judge it on cost per lead. Calls are the first place I'd look, because a call counted on Google often isn't counted the same way on Microsoft unless call tracking is set up on that side too. Check that before you trust any number.

On the 30-conversion rule, I'd add one thing: pick a single primary goal for the first month and make it the event closest to money. For a dental clinic that's a confirmed booking, not a form fill. Imports often bring three or four Google conversion actions across, and the bidder will chase whichever one fires most easily, which is usually a click on a phone link. Set the rest to observation only. That way you can see whether those 12 leads actually turn into patients, instead of just whether the account looks busy.

I'd push back on making the booking the bidding goal for month one. A confirmed booking is the right thing to judge leads by, but Microsoft only sees it if it reaches the account, through either a UET event on a thank-you page or an offline import, and both of those break quietly. At 12 leads a month, the booked subset is probably four or five, which puts the bidder even further below the volume you just said it needs.

I'd keep the lead event as the bidding signal, check that the booking data actually arrives and matches it, and use the lead-to-booking rate to decide whether those leads are worth paying for.

Agreed that the booking shouldn't drive bidding in month one. For the test itself, I'd stick to one lever per window. If you switch the goal, trim the keyword list and fix call tracking all in the same two weeks, you won't know which change moved cost per lead. And at 12 leads a month, a two-week swing is mostly noise anyway. Pick one change, let it run four to six weeks, and only then decide what's next.

Agreed on one lever. At 12 leads a month, a four-to-six week window only gives you 20 or 25 leads, which is still too noisy to read on its own, so don't compare it to last month. Compare it to the Google account over the same dates. If the clinic keeps Google running at the same budget, both accounts see the same season, the same weekday mix and the same competitor moves. Any gap in cost per lead then lines up much more closely with what the Microsoft change did. That turns the problem into "did Microsoft beat Google in the same weeks," which is a lot easier to answer than "did the number go down."

Same dates is the right control, but check that both accounts count conversions the same way before reading any gap. Google's "every" versus "one" conversion setting, plus each platform's click-through and view windows, can shift the lead count a lot on identical clicks. At 12 leads a month, a few double-counted form submissions can flip the answer. Put the same conversion action, lookback window and counting rule side by side, then compare.

I'd also pull impressions and CPC next to cost per lead. If Microsoft's search volume is a fraction of Google's, a cheaper lead can just mean it entered fewer auctions. Cost per lead alongside impression share tells you whether the Microsoft change actually worked or the market just got quieter that month.

Impression share doesn't really tell you whether the market moved. It's the share of eligible auctions you won, so in a quieter month your raw impressions can drop while your share stays flat. And a high share doesn't prove demand held steady either.

To separate a quieter market from a change on your side, compare raw impressions on both platforms over the same dates. If both fell together, the market moved. If only Microsoft's dropped, the change probably cut its reach.