Microsoft Ads at a fraction of Google volume: separate budget or just import?

Most Microsoft Ads advice treats it as a free copy of Google, and that only holds up until you look at volume. In a lot of markets Bing gets somewhere around 5 to 15% of the search volume Google does, and the audience skews older and more desktop-heavy. Take a home services company spending 3k a month on Google. Does it make sense to give Microsoft its own small budget and bids, or import the campaigns and let them run on the same settings? I lean toward a separate budget with a manual check on cost per conversion after a few weeks, because the import tends to carry Google bids into a market that converts differently. I'd like to hear from people who ran this with real volume, especially in B2B, where LinkedIn profile targeting in Microsoft's audience options might change the math.

Separate budget makes sense to me too, but I'd audit conversion tracking before you compare cost per conversion. The import brings over your Google conversion goals, and the Microsoft UET tag often fires on a different event or counts the thank-you page twice. Make sure the Microsoft goal fires once per real lead and is set as primary. On a small account, say home services with 40 to 60 Bing conversions a month, a single double count can swing CPA more than the bid difference you're trying to measure.

I'm not sure the double count is the bigger risk. Tracking is worth fixing, but at 40 to 60 conversions a month, a few weeks of Bing data is thin. CPA can swing by a third from random variation alone, and you're comparing against Google's numbers, which isn't a controlled test. So even with clean tracking, a manual check after a few weeks will mostly show noise. Decide up front what gap would actually change your mind, say 30% on CPA, and give it a longer window before you act on it.

Setting the threshold up front is right, and I'd tie the wait to conversion count instead of calendar time. Random noise in CPA shrinks roughly with the square root of conversions, so at 30 conversions the spread is around 18%, and a 30% gap is worth taking seriously. Below that, call the result "too early." At 40 to 60 conversions a month, that works out to about a month of data. Also judge Bing against its own target CPA, not Google's actual number, because Google's figure comes from a different auction and audience and can't act as a control.

Tying the wait to conversions makes sense. I'd also pull the comparison window in full weeks. Bing skews desktop, so traffic usually dips on weekends, and a window that starts mid-week can tilt CPA in either direction.

Full weeks are the right default, Kai, but the reason isn't the weekend dip in volume. Fewer clicks on Saturday don't move CPA by themselves. CPA is spend divided by conversions, so it only tilts if conversion rate or cost per click shifts too. For home services it often shifts the other way. Someone searching for an emergency plumber on a Sunday evening, on a desktop at home, is usually closer to booking than a Tuesday browser comparing quotes, so weekend CPA could easily come out lower, not higher.

You can check this before picking a window. Pull clicks, conversions and spend by day of week for the last eight weeks. If conversion rate is flat across the week, the start day barely matters once the window is whole weeks. If weekends convert clearly better, a four-week window is still too short to judge a 30% gap, and you'll want both periods to contain the same number of weekend days.