Microsoft Audience Ads on a small budget: worth it or just extra spend?

Bing gets a fraction of Google's volume and its users skew older and more desktop-heavy, so a lot of people treat Microsoft as a free add-on. I'm not sure that holds once audience campaigns get added. In-market and custom audiences look cheap on paper, but on a small budget they can eat through impressions faster than search does, and you end up with a few hundred clicks that tell you almost nothing about conversion rate. Take a local services account spending 300 a month on Bing: the search side might be healthy while the audience campaign quietly burns money on low-intent clicks. My view is to run Bing search as its own clean campaign and only bring in audiences once search has enough conversions to show what a good customer looks like. Has anyone seen audience ads pay off at that budget?

Agree with Theo on the sequence, but I'd make the trigger a number. Add audiences to your search campaign in observation mode first, and only switch them to targeting once each segment has around 30 conversions to compare. Below that, the conversion rate you're seeing is mostly noise.

I'd keep the 30-conversion rule, but how fast you hit it really comes down to monthly volume. At 300 clicks a month, a local services account might get four or five leads, and split across two or three audience segments, it could take a year before any one of them reaches 30. At that pace the rule just means you never switch anything, which is its own kind of waste. With low volume, pick one or two segments that look most like your best customers, run them in observation for a fixed number of weeks, and treat what you see as directional rather than proof.

Agreed on the fixed window. For a local services account, I'd count phone calls and form starts as a secondary signal next to the real leads. At four or five leads a month, that's the only way each audience segment gets enough data points to compare. Calls often come from people further down the funnel, so they're a fair proxy. Just don't let the proxy replace the lead count when you decide what to keep, and label it separately in your reporting so the two don't get mixed up later.

Keeping calls separate from leads makes sense, and I'd make that split in the platform too, not just in your spreadsheet. If calls and form starts are primary conversion goals in Microsoft Ads, the audience optimization will drift toward whichever segment produces the cheapest calls, which is exactly the low-intent traffic Theo was worried about. Set them as secondary goals so they still show up in reporting without steering bids or segment decisions.

Also check how the form-start event fires. If it triggers on first field focus, bounces and bot clicks inflate the count fast, and at your volume that can double the proxy number. Firing it after a real field completion gives you a much cleaner signal.

Agreed on the form-start fix. I'd also set a stop rule before launch: let a segment with zero leads spend up to about twice your target cost per lead, then pause it. At four or five leads a month you won't get a clean statistical answer anyway, so without a limit set in advance it's easy to keep explaining away a dead segment for months. A spend cap makes the call for you and keeps the fixed observation window from turning into an open-ended test.

I don't think a 2x target CPL cap works at this budget. Say your target CPL is 60. A segment can burn 120 with zero leads, and with two or three segments on 300 a month, one dead segment could eat most of the budget before you pause it. A cap tied to a share of total spend, something like 15 to 20 percent per segment, keeps a failed test cheap. The CPL-based cap makes more sense once you have enough volume to trust the target itself.