Written by Igor Ivitskiy
Something in the business is underperforming, the advertising is the most visible moving part, so the advertising is what gets changed. New copy, new targeting, a different bidding setting. Sometimes it works. More often the numbers move slightly and settle back, and three months later the same conversation happens again.
The reason is usually that the constraint was somewhere else, and advertising was simply the part you could reach.
The constraint moves, so find it before you spend
Eliyahu Goldratt’s The Goal makes this point with a scout troop on a hike. The whole group can only move at the pace of the slowest boy, so the only intervention that matters is the one applied to him. Everything else is motion without progress.
Advertising accounts behave the same way, and the slow boy is not always in the same place. I have seen businesses where traffic and conversion rate were both excellent and the real limit was average order value and repeat purchase. I have also seen the reverse: a business squeezing $25,000 of profit out of two clicks and a single order, held back by having almost no traffic at all.
Those two need opposite decisions. The first should not be buying more clicks. The second should be buying almost nothing else. If you do not know which one you are, changing the ads is a coin flip.
Most often, the ads are not where the problem lives
After a long run of these reviews I ended up saying it as a rule: the cause of advertising inefficiency is usually not in the advertising, it is on the landing page.
The traffic arrives, the page fails to convert it, and every symptom of that failure shows up inside the ad account as expensive clicks and a poor cost per acquisition. So the account gets optimised, which is treating the place the symptom is displayed rather than the place the loss happens.
There is a quick sanity check that costs nothing. Look at what an ad promises and then look at what the page says in its first screen. If a visitor has to work out for themselves that they are in the right place, you are paying for that gap on every single click.
Do not automate on numbers you cannot trust
Before adopting any automated bidding strategy, I ask one question: can I trust the recorded conversions at least 80 percent of the time?
If the answer is yes, then the automation has something real to optimise against, and the next move is to remove the soft events from it. Clicks on a phone number, clicks on a button, page views. Those inflate the count and teach the system to buy the wrong people.
If the answer is no, nothing further should be automated yet. Fix the measurement first, write down the date you fixed it, and treat that date as the line from which any later comparison is honest. Skipping this is how businesses end up with a year of reports that cannot be compared with each other.
The budget question has an arithmetic answer
Owners usually ask what budget to start with, and expect a judgement call. It is closer to arithmetic.
Automated systems do not learn from impressions or clicks. They learn from conversions, and they need roughly thirty of them to have anything to learn from. So the working floor is your cost per conversion multiplied by thirty for a month, and fifty is more comfortable.
If a conversion costs you $200, a $6,000 monthly budget is the entry point rather than an ambition. Below that, the system never accumulates enough signal to improve, and you are paying for a learning phase that never finishes. That answer is often unwelcome, and it is better received before the money is spent than after.
What to do with this
Take the four in order, before touching a single campaign setting. Where is the constraint right now: traffic, conversion, order value, or repeat purchase? Does the landing page deliver what the ad promised? Can you trust eight out of ten of your recorded conversions? Does your budget clear the arithmetic floor for the thing you are selling?
Most of the accounts I am asked to look at fail at least one of those four, and none of the four is fixed by rewriting an ad.
Author Bio:
Igor Ivitskiy, PhD, is the founder of Doctor Ads, a mathematician, and has worked on the profitability of paid advertising since 2006. He was ranked #6 in PPCsurvey.com’s Top 50 Most Influential PPC Experts of 2026.