When to Pivot vs When to Hold
Jun 14, 2026Campaign launches. Week one, CAC is $420. Client panics. "Wrong audience. Change the targeting." We change. Week one of new targeting, still $420. "Wrong creative. Different angle." New creative runs. Still doesn't work. Three weeks. Four pivots. Eight grand spent. Zero learning. You know what the problem was? He didn't have two weeks of data. Two weeks is noise. It's not a pattern. Patterns take four weeks.
The noise vs signal rule
Week one data is random. It's affected by platform learning. It's affected by audience size. It's affected by whether you ran at 2am or 2pm. It's not reliable.
Week two data is slightly less random. You've got two data points. That's not a pattern.
Week three and four? Now you've got a pattern. Four weeks of data tells you if something actually works. Not whether it might work. Whether it actually does.
So here's the rule: two weeks, you gather data. Three weeks, you analyse. Week four, if the trend is negative, you pivot. If it's positive, you hold and optimise.
When to pivot
Pivot when CAC is unprofitable and not improving. After four weeks, if CAC is above target and trending sideways or up, something's wrong. Audience might be wrong. Creative isn't resonating. Offer doesn't match messaging. You test each link. If nothing helps, you change the audience or creative.
Pivot when conversion rate is flat despite optimisations. You've tightened the landing page, improved the copy, tested different headlines. Conversion rate is still 2 percent. Probably not a landing page problem anymore. Might be an offer problem. Might be a positioning problem. You pivot the offer.
Pivot when targeting data says you're reaching the wrong people. Your analytics show clicks coming from the wrong demographic. Wrong income. Wrong interest. You're reaching the wrong audience. Pivot targeting.
Real example: Coach ran ads targeting "entrepreneurs." CAC was terrible. Wide audience. After four weeks, we looked at who actually converted. All were business coaches. So we narrowed targeting to "business coaches with 50K revenue." Suddenly it worked. CAC dropped 40 percent. Same offer. Same creative. Better audience.
When to hold
Hold when data trends positive. Week one $1200 CAC. Week two $980. Week three $740. Week four $650. It's improving. Hold. Let it stabilize. Don't kill a campaign that's working towards your target.
Hold when you have high LTV. CAC looks high at $400. But your customer stays 24 months and spends 8000. CAC of 400 is cheap. Hold.
Hold when you haven't hit four weeks yet. This is the biggest one. Most people kill campaigns at week two. By week four they would've worked. But week two looked bad so they panicked.
Hold when you're in platform learning phase. Facebook, Google, TikTokā€”they all have learning phases. First 50 conversions are inefficient. After 50, the algorithm understands. If you're at 30 conversions, hold. Let the algorithm learn.
The patience advantage
Most people quit too early. They have 80 percent of a good campaign and kill it because week two wasn't perfect. You hold. By week four it's working. You have competitive advantage.
Panic kills good campaigns faster than bad ads. The discipline to hold for four weeks separates people who grow from people who spin their wheels.
PS: If you kill a campaign before day 30, you probably should've held longer.