Igor Ivitskiy shows why doubling your Google Ads budget won't double sales. Average ROAS hides losing incremental dollars. He walks through a testing method: raise budget 15-25% per cycle, measure what each extra dollar returns, stop when it hits breakeven, then reallocate.
If you doubled your Google Ads budget tomorrow, your sales would not double. In the next 7 minutes, I'll show you the exact math that tells you how much budget your account can take before it starts burning money. And I'll show you why the number most owners defend, their return on ad spend, is the wrong number to make this decision with. Watch this. Look at this screen. An account spends $10,000 a month and gets $50,000 back. That's a 5x return on ad spend. 5 to 1. Every dashboard in the industry treats that as a healthy account. At a 50% margin, that account makes $15,000 in profit. Now, here's the belief that keeps accounts small. Most owners think this: My return is 5 to 1, so I have to protect that number, and if I spend more, the number goes down. That belief is not stupid. I understand where it comes from. A return target is a simple rule, and it protects you when your data is noisy. It has kept a lot of businesses out of trouble. But it answers the wrong question. Return on ad spend is an Average of money you already spent. A budget decision is a decision about money you have not spent yet. Those are 2 different numbers. So let me show you what happens when you separate them. Look at the second screen. Same account, 3 budget levels. At $10,000, the return is 5 to 1, and the profit is $15,000. At $25,000, the return drops To 3.5. And this is where most people stop reading. But the profit goes up to $18,750. A worse ratio, almost $4,000 more money in the bank. Then look at the third column. $40,000 in spend. The return is 2.6. Still well above the breakeven. Still a number you would defend in a meeting. And the profit falls to $12,750, below where you started. So what actually changed? Look at this next screen. This is the only thing that matters when you scale. Going from $10,000 to $25,000, the extra $15,000 produced 2.5 times its money. Going from $25,000 to $40,000, the extra $15,000 produced Only 1.2 times. That's the whole story. The first step cleared the bar. The second one did not. And here is the part that should worry anyone who reports on return on ad spend. At $40,000, the average is still 2.6. It looks fine. It looks approvable. The profitable early dollars are paying for the losing late ones inside the same average. And the average hides it. 19 years of pulling accounts apart, and this is still one of the most expensive patterns I find. It costs money in both directions. Protect the ratio and you stop short of the peak. Chase revenue and you sail past it. So where is the bar? It's on the screen, and it's one line of arithmetic. Break-even return equals 1 divided by your contribution margin. If $0.50 of every dollar of revenue is your after cost of goods, your breakeven is 2. If it's $0.33, your breakeven is 3. That's the number the next dollar has to beat. Not your best month, not an industry benchmark, not what your agency reported last quarter. Your margin. One division. Now, there's one more thing you need before you touch anything. And it's on the last screen. You might be thinking, fine, I'll just set a target return and let the bidding handle it. Look at what actually arrives. In my data, only 27% of campaigns with a target return actually met it. The median campaign delivered 0.86 of what it was asked for. So, a target is an input, not a promise. If your economics need 3, ask for about 3.5 and check what came back. Here's what to do this week. Take one campaign, work out your breakeven return with that one division. Then, raise the budget by 15-25%. Hold it for a full sales cycle. Then, look at One thing only. What the extra spend brought in divided by an extra spend. Not the account average, the extra. If that number is above your breakeven, do it again. When it gets close to breakeven, stop and move the money somewhere else. That's the whole method. It takes 3 cycles to find your ceiling. And then you know it instead of guessing. The first time I ran this properly on a large account, its real ceiling was almost double where everyone had settled. That's how common this is. The complete worked case is written up on LinkedIn, link below. And if you want the math instead of the hype, subscribe. I take one account pattern apart every week.
