Picture a board meeting at a surgical specialty group a few weeks after their first-ever Google Ads campaign goes live.
The marketing update comes up. One of the surgeons pulls out his iPad, types in the group's name and turns the screen around.
"I don't see us."
A couple of other people pull out their phones and try it. Nothing. Now the room has a theory, and the theory is that the ads don't work.
Here's what was actually going on. They'd spent about $350. Total. That's not a campaign. That's a nice dinner. The card on file got swapped right after launch, so the ads barely ran for the first few weeks. And of the people who did land on the new pages, about one in five clicked to schedule online or call to book an appointment. For a brand-new account, I'll take that all day.
None of that matters once a doctor has done his own test on an iPad.
I love this story, it's more common than you'd think, and it's kind of nuts that it still happens. Advertising works. People have been doing it for a few thousand years. But every new campaign still has to survive its first board meeting.
Googling Yourself Tests Almost Nothing
Here's why the surgeon came up empty. A new campaign on a few hundred bucks is picking its spots, saving its money for people who might actually book. And those people aren't typing in the practice's name. They're typing "shoulder pain specialist near me."
Bonus: search for your own ad enough without clicking and Google stops showing it to you. Your most curious board member is teaching Google to hide the ad from him.
If you actually want to know whether people are seeing your ads, ask for impression share. It's just how often your ad showed when it could have, and it's already in the account. No iPad required.
The Numbers Take 90 Days. Boards Ask at Day 14.
A new campaign spends its first month or so learning who converts. Google's own guidance is to judge performance over at least 30 conversions, such as a month or longer (Google Ads Help). In healthcare, add the time it takes a lead to become a booked appointment, and then a patient. Call it about 90 days before the numbers really say anything.
Most boards I know start asking at about day 14. Some sooner.
Killing something before it gets a fair shot is the obvious cost. But what really gets you is that every pivot resets the clock. New campaign, new targeting, new ramp, and you're back to "too early to tell."
Week Six Is Too Early to Call It
I sat in a midpoint review recently for a specialty platform that grows through physician referrals. The ads were there to give its sales team some air cover in the field. We were about six weeks into a 12-week pilot.
The early signs were good. The campaign started out missing about 80% of the auctions it qualified for, and a few tweaks got it to around 45%. Search click-through was a little over 10%, roughly double what we usually see.
The referrals were mixed. One market went from 5 to 9 a month. Another dropped from 16 to 11. Total for the month was 35 against a baseline of about 40. And a new referral takes around a month to become a patient, so that month wasn't fully baked yet.
Mixed early numbers, in other words. Normal ones.
The client's marketing lead was straight with us. His worry wasn't the data. It was the CFO, who was going to compare roughly $90K over three months against hiring one more sales rep. Fair question. I'd ask it too. You just can't answer it at week six. We suggested giving it about three more weeks before deciding whether to pivot.
And the reason it was so hard to answer at week six is partly on us. We'd agreed on what success looked like at the finish line: total referral lift across the markets, since you can't trace one referral back to one ad. What we hadn't agreed on was what "working" looks like at week six. So the only thing anyone had to go on was a feeling.
And a feeling usually loses to a sales rep.
Agree on the Scoreboard Before You Launch
One of our behavioral health clients asked for this before we'd spent a dime: early signs at 30, 60 and 90 days that both teams sign off on. Smart. I wish every client did that.
Here's what I'd put on it for a patient-facing campaign:
- Day 30: Are the right people showing up?
Click-through rate, how often your ads show when they could, clicks to call or schedule, and a listen to the calls. Are they new patients, existing patients or job seekers?
- Day 60: Are they turning into appointment requests?
New-patient calls and forms, and cost per new-patient call, by market.
- Day 90: Is it filling the schedule?
Total booked appointments and blended cost per booked appointment, compared with your baseline.
- Pull the baseline first.
Twelve months of appointment data by location and service line. "Up" means nothing if nobody knows where you started.
- Write down what would make you stop.
If day 60 shows no new-patient calls, we change it. Boards are a lot more patient when they can see you've already agreed on when to give up.
If you're running a referral campaign like that pilot, same idea, different checkpoints. At day 30, are you reaching the right offices? Look for calls from referring practices' care teams. At day 60, is the field noticing? Referral form downloads, inquiries from new offices, and what your reps are hearing when they walk in. At day 90, compare referral volume in ad markets against your baseline and against markets without ads.
Tell the Board the Rules Before the Game
Most of this is just timing. Do it before launch, not after the iPad comes out.
Send the board a one-page note the week the campaign goes live. What we're running, who it's for, what we'll show you at 30, 60 and 90 days, and what would make us stop. Throw in one line asking them not to Google themselves. It gets a laugh, and it saves you a meeting.
The Monday Morning Version
Find your next launch. Whatever campaign is going live next, write the 30, 60 and 90 day scoreboard now.
Write down the stop line. One sentence: what result, by what day, means we change course. That's the part that buys you the patience.
Ask for the baseline. Twelve months of booked appointments by location and service line. If that's hard to get, that's your first project.
Send the one-pager. Before the next board meeting, not after it.
Your board isn't wrong to ask if it's working. They're just asking in week two a question that gets answered in month three.