If your Bing campaigns spend money and produce nothing, start with the Google import, not the keywords. Someone clicked Import from Google Ads, watched the green checkmarks appear, and left the account alone. I used to do exactly that for clients and call the resulting spend free incremental volume.

It was not free when the imported settings were wrong. Microsoft copied the campaign structure, but tracking, networks, locations and scheduled sync needed their own inspection. A keyword audit cannot fix a campaign that buys traffic outside your service area or counts the same sale twice. Audit the import first. You can find the likely leaks in a first pass of about an hour, though confirming a tracking fix may take longer. Work in this order: stop anything that can overwrite your changes, verify conversions, check where ads appear, then examine queries, bids and Shopping.

Start here: is this still a Google import?

Go to Import > History in Microsoft Ads, then check Import > Scheduled. Past imports tell you how the account began; a scheduled import tells you whether Google can still change it. Other clues include campaign names identical to Google’s, paused Google campaigns mysteriously live on Bing, and Quality Scores that look persuasive only because someone remembers the Google version. Scores do not carry over. An 8/10 in Google proves nothing about this account, and paused Google campaigns reactivating in Microsoft can signal that scheduled auto-import is still syncing status.

Pause scheduled imports before making account changes. Auto-import can sync campaign status, bids, bid strategies and budgets. Leave it running and a later sync can undo manual Bing bid changes or restore settings you meant to remove. I learned this the embarrassing way: clean up an account on Tuesday, watch it revert on Friday. Once the audit is stable, decide what, if anything, deserves to stay synced. Until then, stop the account from arguing with you.

Step 1: Do your UET tag and conversion goals work?

Open Tools > Conversion tracking > UET tag before touching a keyword. The site-wide UET tag needs to be in place for its goals to fire; define the relevant destination URL, event, call or offline goals separately. In imported accounts, I look for an absent UET tag, an Unverified or Tag inactive status, a wrong ID, a duplicate installation, or consent settings that prevent the tag from firing. Use the UET Tag Helper to check the installation, and allow 24 hours for validation. Do not mistake a newly installed tag for a confirmed conversion setup.

Next, inspect every goal that counts toward the campaign’s conversion column. An import can leave a Google goal and a UET goal counting the same purchase. Then one sale appears twice and automated bidding optimizes against a number that is not real. Keep one primary goal per action, exclude the duplicate from bidding, and turn off conversion-goal import if it would bring the duplicate back. Compare Microsoft Ads conversions with orders or CRM records for the same action and period.

If Microsoft reports more conversions than your orders or CRM, stop here. Network and bid decisions built on a broken count are not decisions; they are guesses with a dashboard attached. Get the count credible before asking an automated strategy to use it.

Step 2: Where is your spend actually showing?

Check Settings > Ad distribution at campaign level. Search campaigns can be extended into Audience Network, putting search ads on MSN, Outlook and Edge placements with display-style intent. That is a different job from answering a search query. One operator breakdown puts the waste at 30% or more of budget with low click-through. The old negative-bid-modifier fix no longer works; excluding this traffic requires a manual support request and unchecking account-level image and multimedia auto-apply. For a lead-generation account, opt out, then watch CPA for two weeks before judging the change. Do not lump the result together with search traffic and call it a keyword verdict.

Then run Reports > Website URL (publisher) for the last 60 days, segmented by campaign. Sort by spend and find publishers with zero conversions. That is your search-partner leak list, not a reason to condemn every partner at once. Exclude the worst offenders, then test one campaign with partner traffic off entirely. Name the tradeoff before you click: disabling partners also removes DuckDuckGo, whose ads serve through Microsoft’s syndicated partner network. For most lead accounts I have seen, that tradeoff pays, but use your publisher report to make the call.

If a large share of spend never touched Bing.com, you do not yet have a keyword problem. First find out what inventory bought the clicks.

Step 3: Are you buying clicks from places you do not serve?

Microsoft’s default location intent is People in, searching for, or viewing pages about your location, not physical presence alone. The second clause matters to a local advertiser. Someone in another state researching your city can see the ad, click, and never become a job. For a fixed service area, switch to People in only. Location exclusions alone do not settle the issue when the campaign is still eligible to buy interest in a place rather than presence there.

Check the setting below campaign level, too. Microsoft lets you set location intent at ad-group level; where it is set, it replaces campaign locations rather than refining them. I have opened accounts where half the ad groups targeted the right city and the other half quietly targeted the whole country. Work campaign by campaign, then inspect ad groups for stragglers. An account-level glance will not catch a single ad group spending outside the service area.

Match the location rule to where you can actually sell. If the search-terms report keeps surfacing cities you do not serve, return to this setting before adding another pile of negatives.

Step 4: Why are the wrong queries matching?

Pull a 90-day search-terms report and sort by spend, not clicks. Bing’s query pool is smaller than Google’s, so an imported broad keyword can gather a spread of near-misses that did not look costly in the Google account. Close variants deserve a close read here. I used to tell clients broad match was safe if the intent looked close. I was wrong on Bing: without rebuilt controls, the loose terms can burn through a small pool fast.

Now inspect the negatives you assumed survived the import. Microsoft Ads has no broad-match negatives, only phrase and exact. Imported broad negatives arrive as phrase, which can leave gaps unless you add both word orders. Look for the terms you thought you had blocked years ago: job seekers, DIY searches and competitor research. The fix is mechanical, not glamorous:

  1. Find the highest-spend irrelevant queries and review the negatives meant to block them.
  2. Rebuild the top 20 negatives by hand in phrase and exact, including reversed word orders where needed.
  3. Set aside 15 minutes each week for a negative-keyword pass during the first month.

If your highest-spend search term has zero conversions, pause the keyword before rewriting the ad. Better copy for a query you did not want is still better copy for the wrong query.

Step 5: Are your bids and budgets set for Bing or Google?

Imported bids can miss in both directions. Google-level CPCs may overbid in Microsoft’s cheaper auction, while imported Target CPA or Target ROAS targets can starve on Bing’s smaller volume. The thresholds I use are 30 or more conversions in 30 days before Target CPA and 50 or more with values before Target ROAS. Below those levels, I start with Maximize Conversions uncapped, establish that clean, measured volume exists, and tighten only then. None of that works if Step 1 left conversions untrustworthy.

Budget deserves the same separate treatment. A copied $200/day Google budget does not make a Bing campaign that can spend only $30 profitably into a scale opportunity. If a campaign runs into its cap by mid-morning, inspect what it is buying before raising the limit. If it shows below-first-page warnings, inspect the imported bids and the queries they are supposed to win. Both warnings demand diagnosis, not a reflexive increase.

Check device and audience splits before adjusting targets. Bing still skews desktop, especially for B2B, so an imported -100% desktop modifier from a mobile-heavy Google account can quietly kill useful traffic. Imported in-market lists and bid adjustments can also stack against smaller volume until little serves. Reset device and audience modifiers to zero, let clean data accumulate for two weeks, and make the next adjustment from what Bing actually did. If a bid strategy has been in Limited Learning for 30 days, do not assume day 31 will fix the volume problem.

Step 6: If you sell products, is the feed healthy?

Shopping on Microsoft rarely fails on keywords. It can fail in Merchant Center while campaigns continue spending on whichever products remain eligible. Open Merchant Center > Diagnostics and sort by products affected. Check for an unverified domain, missing GTINs, price mismatches between the feed and landing page, and unsupported categories. Verify the domain first, work down from the biggest affected product group, then re-fetch the feed. That order gets more useful products back into consideration before you debate a bid adjustment.

I have seen catalogs where 40% of SKUs stayed disapproved for months while the Shopping campaign optimized around the remaining 60% and called it performance. Keep campaign structure aligned with the feed so removed products do not leave empty product groups behind. Use the Shopping Dashboard and Feed Audit to confirm health before judging ROAS. A campaign cannot prove demand for products it was never allowed to show.

No bid tweak rescues a feed where your best sellers are ineligible. Fix eligibility, then evaluate what the campaigns did with it.

What should change in the first 14 days?

A proper takeover does not begin with new ads. It begins by stopping the account from undoing its own fixes, then earning the right to spend against trustworthy conversions. If you hire someone to take over a Bing account that spends without converting, this is the sequence I would hold them to. It also matches the reason groas runs Search Ads across Google, Bing and ChatGPT as separate channels, with bids and budgets worked continuously rather than one import left to drift.

  1. Days 0–2: stop the overwrite and check measurement. Pause scheduled Google import. Confirm the UET tag fires on key page templates, and keep one primary conversion goal per action. Compare reported conversions with the records that tell you whether a sale or lead happened. If they do not agree, do not hurry on to bidding.
  2. Days 3–7: close distribution and location leaks. Opt search campaigns out of Audience Network, exclude the worst syndicated partners identified in the publisher report, and switch fixed-service-area campaigns to People in only. Rebuild the top negatives in phrase and exact, with both word orders where needed. This is the part where the search-terms report starts describing the business you actually run.
  3. Days 8–14: give bidding clean inputs. Reset device and audience modifiers to zero. Move thin converters to Maximize Conversions uncapped only after tracking is sound, and reintroduce Target CPA once a campaign clears 30 conversions in 30 days. Check Shopping feed health before treating its ROAS as a bidding verdict.

The parallel Google Ads audit checklist applies the same discipline to Google. It is useful precisely because it helps you see which Bing settings need their own decisions. If your new manager proposes headlines before tracking and networks are clean, they are decorating a leak.

What should you skip until then? New ad copy, new keywords and more budget. I have watched owners double Bing budgets to force volume while Audience Network took 30 cents of every dollar and the UET setup double-counted the two conversions they got. More spend in that setup buys more of the same traffic and teaches the bid strategy the wrong price for a lead. Fix measurement and distribution first. Let search terms and Microsoft-sized bids determine where to go next.

Your first move today takes ten minutes: open Import > Scheduled and pause it. Then open Tools > Conversion tracking and check whether one primary goal matches your CRM count. If those screens look wrong, you have found where to start; work the rest of the audit in order. If you would rather not babysit sync settings, publisher exclusions and weekly negative passes, that is what groas does inside Search Ads: it treats Bing as its own channel, works bids and budgets continuously, and keeps a human owner accountable for pipeline and CPA. Either way, stop letting a Google copy spend like Bing.