You spend $2,000 a month on Google Ads, qualified calls are erratic, and your agency wants to add Bing. Should you fix Google first or fund a second channel? Smart operators disagree because they have seen opposite failures: expansion that turns one messy account into two, and endless Google repairs that postpone a useful Microsoft test. I think measurement readiness should decide, not Google perfection. But the case for waiting deserves a proper hearing before I sign the test budget.

What the second-channel pitch actually buys

For a first-time buyer, this is practical, not philosophical. You already pay for Google Ads. Someone offers to run Microsoft Ads alongside it, reaching traffic across Bing, Yahoo, and DuckDuckGo. You want to know whether that fixes a lead problem or spreads it around.

The pitch sounds simple because the work sounds similar: keywords, negatives, bids, budgets, town targeting, call tracking. Both platforms buy search clicks for your services and locations, then report which clicks turn into calls or form fills. Neither can tell you what a good lead means until you define it.

Local Scope SEO describes running Microsoft Ads with the same job types, towns, and tracking as its Google work inside one $700 ads plan. That shared structure makes sense. A shared blind spot does not. If Google has no clean record of qualified leads, copying its campaigns will not create one.

The price of adding a channel also varies. Local Scope includes Microsoft with no extra management fee, while ClickedOn lists $1,800 a month for one channel or $2,800 for two or more, plus 12% of spend above $5,000.

Ask what Bing adds to the fee before asking what it adds to performance. Then ask what ongoing work that fee buys. An import button is not an account manager.

Position one: fix Google before adding Bing

The strongest case for waiting is not loyalty to Google. It is refusing to pay for a second experiment when the first one cannot tell you what happened.

Broken tracking makes a second account harder to judge

I used to tell clients an import was basically free expansion. I was wrong. A Google import is a one-time copy by default, not a live sync. Someone still has to maintain both accounts, including their budgets and negatives.

Conversion measurement needs separate attention. Google tags do not report conversions to Microsoft. Importing campaign structure does not mean Microsoft conversion tracking is ready.

SF Digital’s import breakdown flags conversion tracking that does not transfer, bidding changes when no Microsoft goal exists, and scheduled re-imports that can overwrite Microsoft-specific fixes. Those are not cosmetic setup details. They affect what the campaign optimizes toward and whether you can judge the result.

Add a channel before sorting that out and you do not get twice the learning. You get another bill with an explanation attached.

Do not expand a measurement problem. Fix it.

A thin budget can leave both tests inconclusive

Say you spend $2,000 a month on Google and carve out $300 for Bing to “see what happens.” The budget sounds modest. The question is whether it buys enough activity to answer anything.

Trackbee offers a rough planning rule of $500 a month per campaign for a few weeks to reach about 30 conversions for automated bidding. That is a rule of thumb, not a promise that your local service campaign will produce 30 qualified calls.

EverestX recommends 10 to 15% of the Google budget for 60 days, and advises skipping Microsoft below $1,000 in monthly Google spend. The useful point is not that one threshold fits every account. It is that a test needs enough budget and time to produce something you can interpret.

Two starved campaigns can produce two inconclusive results and one confident-sounding report. If the Bing allocation comes out of Google, the trade-off matters there too.

Do not call a token allocation a proper test.

A weak offer can fail on both platforms

Fix-Google-first has its strongest argument here, and it has nothing to do with Google.

If your page closes one in fifty visitors, your quote process takes three days, or your reviews look thin next to the competitor above you, Bing will not repair that. A cheaper click still lands in the same sales process.

In home services, I have learned to look at the call answer rate before treating bids as the whole CPA problem. A campaign cannot book the estimate if nobody picks up.

Before buying another source of traffic, check the parts both channels share:

  • The offer: is there a clear reason to choose you?
  • The page: does it give visitors enough proof to act?
  • The response: does someone answer the call or follow up on the form?

The advocate for waiting has a straightforward case: repair the shared failure before paying to feed it.

A broken close does not become a working close on Bing.

Position two: give Bing a bounded test now

The strongest case for testing starts with a distinction: Google performance measures Google performance. It does not settle what another auction can deliver.

Google still dominates volume, at about 90% global share against about 4% for Microsoft. But the argument for Bing is not that it replaces Google’s volume. It is that a smaller, separate auction may deserve its own budget.

Google cannot establish what Microsoft traffic will do

Microsoft reaches a different audience mix. Microsoft cites 1B+ monthly users and 44M US searchers Google cannot reach, with an audience that skews older, higher-income, desktop-heavy, and often B2B.

There is also a concrete price argument. A SmartSites analysis of 73 advertisers found average search CPC of $1.54 on Microsoft versus $2.93 on Google: 47% less. Of those advertisers, 64 paid less on Microsoft, with a median saving of 59%.

Cheaper clicks do not establish cheaper qualified leads. They do establish a reason to test rather than assume the Google result applies everywhere.

Waiting for Google to hit an arbitrary CPA target tells you nothing about that separate auction. It only delays finding out.

Google’s unfinished work is not evidence that Bing will fail.

Some Google problems belong to Google

A national competitor bidding up your core terms in Los Angeles is not the same problem as nobody answering your phone. Neither is a Performance Max campaign eating brand queries. Those are reasons to work on Google, not automatic reasons to block Microsoft.

EverestX makes audience fit central to its when-to-test framework, highlighting older, higher-income, desktop, and B2B audiences and recommending a 60-to-90-day commitment rather than 30 days.

If your buyers sit at office desks on Windows defaults, “Google still needs work” is not much of a counterargument. Google usually still needs work. That cannot be the permanent veto.

The test advocate is not asking you to stop fixing Google. The argument is that a measurable Microsoft experiment can run while Google-specific repairs continue.

Separate an account problem from a business-wide problem before deciding to wait.

Importing reduces setup work, not responsibility

The cheap way into Microsoft is also the easiest part to oversell. Yes, you can copy Google campaigns. No, that does not mean they are ready to run unchanged.

SF Digital’s import audit flags small-town targets widening to counties or regions, search partners switching on, and scheduled imports overwriting local fixes. Each needs inspection before the test becomes readable.

The ongoing work still includes:

  • Checking Microsoft conversion goals and tracking.
  • Adapting bids and negatives.
  • Reviewing locations and network settings.
  • Protecting Microsoft-specific changes from later imports.

That is the strongest version of the test-now argument: not copy it and hope, but fund a separate experiment and manage it separately.

The import is a starting point. The upkeep is the service.

My verdict: test Bing when measurement is ready

I side with the Bing-test camp for advertisers who can identify qualified leads separately by channel and fund a bounded experiment. Measurement readiness is the deciding consideration. Google does not have to reach some imaginary finish line first.

That does not excuse a broken offer or a budget too thin to learn from. Those are reasons the experiment may not be viable. But “Google isn’t perfect yet” is not a useful stop rule. It can postpone Microsoft indefinitely without telling you anything about Microsoft.

Cartoon of a marketer splitting a tiny ad budget into two thimbles labeled Google and Bing

Wait when you cannot trace a qualified lead

I would hold the test if Microsoft conversion tracking is not working, call tracking cannot separate Bing from Google, or nobody agrees what counts as qualified. I would also hold it if network traffic is mixed together so badly that the search result cannot be judged.

A weak close needs attention too. If nobody answers the phone or the quote process loses interested buyers, fix that before paying for another source of them. That work helps Google as well.

The readiness question is concrete: can you name the outcome you want and see which channel produced it?

If you cannot identify a qualified lead per channel, fix measurement before expansion.

Ring-fence the budget and accept the limits

Once tracking works, the test needs a budget that stays separate from a vague promise to “try Bing.”

On a $2,000 Google budget, I would consider roughly $300 to $500 a month ring-fenced for Microsoft for 60 days. That is a proposed test allocation, not a guarantee of enough conversions. The right decision still depends on whether that money can buy useful evidence in your market.

You may get 12 clicks a week in a small town and prove very little. Accept that possibility before launch. A capped inconclusive test is a known risk; an open-ended inconclusive campaign is an expensive habit.

Nor should spending the allocation become the goal. I covered that failure in Stop Trying to Make Microsoft Ads Spend Your Budget: a test that under-spends while holding CPA provides information. It is not automatically a defeat.

Fund a real experiment, not an obligation to spend.

Put three decisions in writing before the import

This is where joint-channel arrangements drift. Everyone agrees to “test Microsoft,” but nobody agrees what that means.

I want three things written down first:

  1. A spend cap. Set the monthly allocation and require approval before increasing it.
  2. A review date and stop rule. Usually day 60, with the qualified-lead and cost-per-qualified-lead results that justify continuing, stopping, or acknowledging insufficient evidence.
  3. One qualified-lead definition. For example, a booked estimate with a real address, not every form fill.

Without that last sentence, one report celebrates cheap forms while another counts closed jobs. Then everyone argues about the table instead of the business.

The cap limits the downside. The definition makes the comparison usable. The review date stops “testing” from becoming the account’s permanent status.

Agree what success means before anyone spends toward it.

Hire for separate scorecards and a named owner

Any agency managing Google and Microsoft together should show cost per qualified lead separately, then judge both against the same business goal.

Blended CPA can hide a weak Bing test inside strong Google volume. Blended lead counts can hide junk forms inside real booked jobs. A tidy combined number is not transparency if it conceals the decision you need to make.

I would also ask who does the work after the import. Scheduled copies and a monthly glance are not enough. I want bids, negatives, locations, and conversion goals reviewed for Microsoft specifically, with changes logged. I laid out the hiring choice in Bing specialist vs Google agency with Bing tacked on vs one manager for both.

Before signing, ask:

  • Will I see qualified leads separately for Google and Bing, down to campaign and town, or one blended total?
  • Who maintains Microsoft-specific bids, negatives, locations, and tracking, and how often?
  • What result at the review date keeps the test alive, and what stops it?

Shared goals, separate scorecards, named owner. That is the operating model I would buy.

My ruling is to run the bounded Bing test once those conditions are met, rather than wait for Google to look perfect. If qualified leads cannot be measured separately, fix that first. Otherwise, continued Google repairs are not a reason to leave the second auction untested.

That is how groas runs it: one flat monthly fee covering both channels, continuous execution instead of a monthly check, and every change logged against qualified pipeline rather than clicks. Book a consultation with the cap, review date, and lead definition agreed up front. Give Bing a chance to earn its place, not an unlimited budget or an indefinite wait.