I’ve been working for 15 years in a group of companies focused on performance (measurable) marketing / online advertising. I often get asked, “Where can I advertise cheaply?” or “Why is it so expensive to acquire users or customers now?”
This article is not for experts, but anyone who encounters marketing and wants to understand what's going on.
It's time to summarize observations about the market, where it's moving, what point we're at now, and how long this will all continue.
When it was tough but still fine. Before 2014
Let’s go back about 14 years. Mobile internet was just getting started, and desktop computers and laptops still ruled. There was plenty of online advertising, and it was genuinely “performance”:
if someone made a purchase (or became a lead) on your site, you could see where they came from. Sure, people complained about tracking inaccuracies and debated the best methods, but overall, performance ads worked.
Why did it matter?
One of the most significant advantages of online advertising has always been the ability to track what really brings in customers — and then invest only in those sources. This boils down to attribution: identifying who or what influenced the user who bought something.
The most popular model is Last Cookie Win (“whoever touched the customer last, gets the credit”). If someone clicked five different sites with your ads, only the last one would be recorded in your analytics.
Is this system good? Like Churchill said about democracy: “It’s terrible, but the alternatives are worse.” Other models (like “first click”) or even “multifactor attribution” exist, but at scale they’re tricky.
A known case: a large sports corporation, tired of Last Cookie Win, spent millions looking for a “better” attribution model. After all their research, they ended up sticking with Last Cookie Win.
The transition period. 2014-2019
Performance advertising worked, but marketers complained about its flaws. Between 2014 and 2019, smartphones and mobile apps grew wildly, almost like a separate ecosystem from desktops. Specialized software and attribution systems appeared for mobile, but they still followed the LastClick principle. Everyone talked about omnichannel (e.g., researching on a laptop, buying on a smartphone), but it felt like a problem for the future.
Today’s reality. 2019+
I guess the point of reference could be Сovid. 2019. A huge global shift. Even though it didn't change much, it accelerated the already existing trends many times over.
Nobody cared about subtleties like attribution models or the type of performance advertising available. Tectonic shifts were underway. If you offered a remote service, you grew by multiples across all marketing channels.
If you were tied to physical presence, you had all channels flying into the abyss.
Some won, some lost.
No sooner had COVID finished, and the people reflected on what had happened and changed — geo-tensions, wars, and conflicts between countries increased dramatically. From a global economy, began to divide into regional economies. Also big shifts that left smaller changes in the shadows.
The situation is still complicated, but people get used to everything and now it's okay to reflect and look around a bit.
While the world was busy fighting microbes, ideas, and all sorts of enemies:
- Omnichannel marketing won
- TikTok shot into space, with its "versions" popping up in every social network
- ChatGPT and its twin brothers
- Messengers edging closer to super-apps—basically the internet inside themselves
- 5G, high-quality internet, satellite internet. These days, you can only scare kids by saying a phone video over 2MB used to take ten minutes to load.
The actual customer journey today might look something like this:
- Person learns about a product from a friend in a mobile messenger
- Finds reviews on TikTok and YouTube; watches a couple of vids (one is great, one is rubbish, but stats show two views)
- Searches for the product on Google using a laptop
- Asks ChatGPT about it on a phone
- Uses cashback or sends the link from a work device to a home computer
- Made a purchase of goods from home, on the way clicked on some promotion of the store/service that offered a discount coupon.
Analytics might record this entire path as an “organic” visit or say “the loyalty program worked,” even though many steps actually happened. A marketer who paid for a fancy tracking system might see 3 of those 6 steps and decide it’s all “Google” or “SEO traffic.” But here, too, it is not clear whether this is a benefit or not if, in the end, this information only misleads you.
Oh, and that's not to mention the numerous data protection laws, GDPR, etc., at the level of states, browsers, and general platform regulations. All of these are under the banner of data protection, forcing you to click on pop-ups on every site you visit. It's not clear what it protects you from or who it protects, but the convenience hasn't improved.
People don't like ads, so many use tools to remove them (some are even built into browsers).
These are all things that try to hide the traces of the user.
What about Partner networks?
In some of our businesses, we run partner networks: connecting advertisers with bloggers or publishers who can send them customers. If a blogger promotes sneakers and a user buys them, the blogger earns a cut.
We began to notice that despite the large investment in tracking who brought who to whom, the situation was not getting any better. I even wanted to curse advertisers who tried to cheat and reject sales. It happens, too, but more often, it's a different story.
According to their analytics systems, the client came from a search, their mailing list, or contextual advertising, not from the partner. Therefore, the partner's sale is rejected but recorded in the "organic", as if it were their regular client who suddenly decided that he needed another pair of sneakers.
But the reality turned out to be simpler. If we, whose business depends on our ability to see who brought clients to whom from whom, do not see some of the orders, then the situation is no better for advertisers. They themselves do not see which client came from where.
Today, most of the market is working blind. More precisely, they have many beautiful dashboards and plausible interpretations, but there is no reason to believe that they can really see their clients' paths on a scale.
That means that marketing decisions nowadays can't be called data-driven.
And we should not expect fair and effective decisions on where to allocate marketing budgets.
What's there to do?
- Acknowledge the problem. Online ads aren’t as accurately measurable as everyone assumes. The margin of error is huge, and we need to keep that in mind.
- Focus on what can be measured. A good product is always easier to promote, so it makes sense to rely more on the value you give to users, which is quite accurately tracked by current analytics systems.
Yes, advertising will continue getting more expensive. The internet is consolidating into major platforms and AI tools: 100 shops become Amazon, 1000 info-sites become a direct GPT answer.
Having made a very big circle, we're back to a classic quote from the early twentieth century, one of the founding fathers of the advertising industry, John Wanamaker (another version by Ogilvy):
“I know half the money I spend on advertising is wasted. The trouble is, I don’t know which half.”
Maybe that’s not so bad. People have built huge businesses in less measurable eras, and it gave birth to advertising as an art form. The reduced “predictability” might just spark more creativity and innovation. And since advertising is basically about communication, there’s never been a better time to learn how to understand each other and share ideas effectively.
