Interview with Igor Ivitskiy, PhD, Founder, Doctor Ads

I face this all the time in audits. This is one of the main reasons clients come to me for audits. They come with roughly the following request: “Everything is perfect. My agency, my contractor, and my advertising department send perfect reports. Everything is just unbelievably great, but the money isn’t increasing and the business isn’t growing. Find the reason.”

In all these accounts, when you clean them up, the results take off. And the owner is in shock because they spend exactly the same amount on marketing and advertising but get more results.

I had a small online store. Then I opened Google Ads. Of course, I chose general keywords like “microwave,” “electric kettle,” and “hair dryer.” I don’t think I even wrote any ads. I copied them from the catalog pages of my online store. Do you know what my first ads looked like? The headline was something like “All microwaves,” and the ad text said, “A microwave is a device designed for heating food.” They were terrible ads. Of course, I wasted money. I got a lot of clicks, but I didn’t know that you need to look at search queries and that relevance matters. I looked at the keyword “microwave”: a lot of money wasted, a lot of clicks, and not one sale. I decided that Google Ads didn’t work.

Previously, I worked in pure mathematical modeling. Thousands of scientists work in this field all over the planet. There, my approach wasn’t unique; I was one of thousands. Then I moved into marketing. I didn’t betray science; I changed laboratories. I do the same science in the laboratory of business, money, and advertising. I work a lot with small and medium-sized businesses.

At one conference, a guy was giving a talk, and he said, “Our ROAS, the return on the money put into advertising, was 19,000%.”

Honestly, I was shocked. I didn’t know what you would need to sell to achieve a 19,000% ROAS. I thought, “Maybe this is a sweet spot. Maybe I need to go into this field.”

Then he said roughly the following: “We set up advertising for a client, and someone whose click we paid $1 for bought goods. This means we were able to give the client a ROAS of 19,000%.”

The hall just exploded with applause, and after the conference, everyone came up to this guy and asked for his contact information and how to work with him.

I was laughing because how can you attribute a sale to one click? I’m sure that the ad campaign had not just one click, but thousands of clicks, maybe tens of thousands. Most likely, there were 2,000 clicks at $1 each, and you need to compare the profit not to $1, but to the $2,000 spent on the whole campaign. But nobody mentions this.

One of my businesses is an online school. People register for a webinar, and after the webinar, they buy the course. I knew that 40% of registrants attended the webinar and that 6% of attendees bought the course.

Then I started testing additional advertising channels. I launched the updated Google inventory. I think Performance Max had just appeared; this was four or five years ago. It gave me conversions at eight times less than the usual cost, and I felt like I had hit the jackpot. Although I’m a mathematician, I have a major personal flaw: I like to count money I haven’t earned yet. In my mind, I was already buying myself a new car. Later, it turned out that the numbers were misleading.

I should have at least checked attendance. Fewer than 1% of these leads attended the webinar, compared with the usual 40%. I moved the entire budget there. It was only when my profit fell dramatically a month later that I realized something was wrong. I started looking at attendance by channel and saw that these were just dead leads. Then I started counting not just webinar registrations as conversions, but also webinar attendance, email confirmation, or a visit to Telegram or ManyChat after registration. And this worked well.

The company had advertising budgets of $5–6 million a month and was optimizing for conversions. On paper, everything looked good—too good. They had a separate person on staff whose job was to set up and track conversions. The absurd thing was that I went in and looked at his purchase conversions. There were roughly hundreds of thousands.

Then I overheard the owner say, “Tens of thousands of purchases a month.” I said, “Wait, hundreds of thousands or tens of thousands?” The owner said, “If we had hundreds of thousands, I would be the richest man on the planet.” He wasn’t looking at the absolute number of conversions; he mostly looked at relative numbers: the conversion rate and the cost per conversion. He liked everything.

It turned out that this person had set the conversion action to trigger not when a payment was made, but when someone clicked the payment button. There is a saying that there is nothing in the world more permanent than the temporary. He set it up temporarily, forgot about it, and everything kept running for months. When there is a chain of dependencies, someone waits for an answer from someone else and decides to do something temporarily while waiting. Then the answer doesn’t come, or it comes but goes unnoticed. And what was temporary becomes permanent. This is a common problem.

I had a client. I looked at the account, and he said, "Our margin is colossal. We pay $20 for a client in Google and get $1,000 of profit from a client." I said, "Why don't you increase your cost per conversion fivefold—your target cost per acquisition (Target CPA)? Right now, your Target CPA is $20, but you can afford $100." He said, "I can." I said, "What stops you from paying $100 for one lead instead of $20?" He said, "It is expensive." I said, "Expensive for whom? You have $980 of net profit. If you pay $100 to earn the same $1,000, your profit will fall from $980 to $900, so net profit falls by about 8%. And the number of clients will not be 100 but 1,000. Net profit will grow almost tenfold."

A few months later, we had a call with the client, and he said, "The best decision in my business was to just increase the budget and the cost per lead." He had some kind of mental ceiling, a mental wall, that made him think more than $20 was expensive. I just explained to him that if there is room, why not? He said, "We grew so much that now we are the biggest in the market."

Nobody segments by conversion. They look at the number and believe that number, but when you segment, right away there are a lot of questions.

I say, “Look, you have 20. 20 what? Conversions. Which ones?” It is the same as asking, “What is 5 pineapples?” How much is 5 pineapples plus 5 watermelons, plus 5 oranges, plus 5 apples? The person says, “It will be 20.” I say, “20 what?” He says, “Fruits.” I say, “OK. How much is 5 chairs plus 5 shirts, plus 5 cats, plus 5 kilograms of flour?” He says, “It will be 20.” I say, “20 what?” “Kilograms of animals.”