Interview with Kagi
The internet is run by adtech right now, and all the biggest companies are thriving on that, and I think the sooner we get out of that, the better. It is certainly going to be hard, and I'm not against ads per-se, but I'm against ads everywhere, ads by default, ads making our lives crappier and more miserable.
This is a casual interview that I had with Vladimir Prelovac, the founder of Kagi, in March 2022.
DKB: Why did you decide to work on Kagi? Building a new search engine is a pretty ambitious project.
Vladimir:
It was 2018, and I was fresh out of the corporate world after the acquisition of my startup. I wanted a search engine for my kids that I could trust. My main motivation was to prevent my kids from being exposed to ads and tracking from a very young age.
Kids start to use computers at a young age since they need it for school, and 99% of kids start with Google. They’re providing free Chromebooks to schools, and it’s convenient to use Google, so that’s what kids are using.
I didn’t like the idea of them growing up with ad supported technology that would influence them from a young age – shaping their thoughts and views of the world.
Back then, there was literally no search engine like that, because DuckDuckGo was also ad-supported. I was paying for YouTube Premium because I didn’t want my kids exposed to ads, so I started thinking about why there wasn’t also a search engine without ads that you could pay for. The revenue would come from the users, perfectly aligning all the incentives.
And that’s how Kagi got started.
I wasn’t alone. With significant funds from my previous exit, I was able to bootstrap the entire operation with around 10 people.
DKB: I want to talk more about your business model. You seem convinced that people will pay fo a search engine, even though there are no good examples of this in the past.
What makes you think that people will pay for something they’re used to getting for free?
And will the pricing be unlimited searches per month, or some kind of per query pricing?
Vladimir:
Before I tell you my answer, I’m curious to hear what you think about all that, and what your view is as someone in the field?
DKB: People pay for YouTube Premium, so there’s certainly some subset of people who would pay to avoid ads. I’m not sure what the percentage is, but I would imagine it’s a relatively small percent of users that do this.
When it comes to how pricing should work, I think pricing per query is scary. Users would be scared to do queries because they have to mentally calculate whether it’s worth the 2 cents or whatever they’re paying for it. An unlimited pricing model would probably put people’s mind at ease.
I know Listen Notes, a podcast search engine, has their main search for free but charges for a more advanced search with more filters. That’s another interesting way of monetizing where you have a simple free version, and a more advanced paid version for the power users.
Vladimir:
You’re spot on. Paying for a search engine is unusual, but that’s the only choice if you want to build an independent search engine that isn’t ad supported. You could try out donations, but I don’t think that would be dependable.
I think we have a significant tailwind in the world today with privacy being a hot topic. People are waking up to how their information is bought and sold online. Search engines have also been criticized for introducing biases and agendas. In my opinion, the job of a true web search engine is not to be the arbiter of truth. It should provide information based on search relevancy algorithms, not based on the social, political, religious, or whatever views its leadership may have. The only agenda it should have is to provide the best possible customer experience.
For all the big search engines, the issue is that their customers are advertisers, not the users. This creates all sorts of conflicts of interest. With Kagi the users are also the customers, and this means we can design a better search experience, and we hear from a lot of users that such an experience is worth paying for.
So when you think about it, you're quite right. Five billion people use a search engine, probably 99% would never pay. The thing is, that tiny minority of people who think differently, the 1% that would pay, is still quite a large number. That's 50 million people.
YouTube Premium has 50 million subscribers, even though everyone can watch YouTube completely free! And yes there is of course Google Search, you can use it for free, but you're exposed to ads and tracking and your productivity takes a hit because of a poor user experience, optimized for serving many ads.
You can use an ad blocker with YouTube to remove ads, and despite that, there are still 50 million people paying for it. This shows how much these people do not want to be exposed to ad tech.
Kagi doesn't need a lot of users, because as I said, we are not a Google killer. Google has a huge moat, and a 20 year headstart, and the web is now much harder to crawl than it was 20 years ago, so we will probably never have Google scale, but we don't need it.
We are trying to build a sustainable business for a small subset of users who think differently. We just need a couple of tens of thousands of such users to be sustainable, and that's our first milestone.
When we are talking about sustainability, we should note that searching has a cost. We are not used to that cost right now because advertisers subsidize it, but if you are doing a search, it has a real cost, and we are trying to bring that cost as low as possible.
It would be about one cent per search or something like that, but still for many users the idea of paying for search is frightening. I like to compare it with electricity, do you think of cents being spent when you have a light on, or do you think about the convenience of having the light on?
The challenge we are facing can be described through this analogy: imagine electricity was completely free for the last 20 years, but there was a catch, you had to have ads on your walls 24/7. And now someone comes along and says we are going to offer electricity at a few cents per hour, but we will completely remove all the ads from your home.
The idea of paying for a utility that was free for so long would be unthinkable for most, but this is what we already have in reality and everyone is okay with it. This is why I think the idea of paid search also makes sense.
Twenty years of ad-supported search have created resistance to the idea of paid search, but a lot of people are already coming to the realization that the predicament they have found themselves in with ad-supported search is not good for them.
DKB: Where do the results for Kagi come from?
Vladimir:
We get results from Google, Bing, and our own non-commercial indices. We have two of our own indices, one for non-commercial web search, and another for non-commercial news.
The non-commercial indices can have a big impact on the quality of results since it helps us avoid spam, SEO optimized content, and pages riddled with ads. It also allows us to get expert opinions from small websites and personal blogs.
I think a higher quality of search is ultimately worth the price, because you can find something quicker with just one query, and we can save you half a minute per search, especially when you're doing research.
The diversity and quality of Kagi results is really unique on the market, and on top of that we have pretty cool features. For example, you can block or promote websites in your results, so if you have favorite ones or if there are spammy sites you don't like, you can modify your results.
We also have the lenses feature where you can create a lens, which is basically a subset of up to 10 sites. So you can create a lens for programming or recipes with your favorite sites.
DKB: The customization features on Kagi are quite powerful.
Do you see your target audience as the kind of person who would want to dive into all of these settings? Or is Kagi going to have good settings by default for someone who just wants to search without customizing anything?
Vladimir:
We have reasonable defaults, and for most people Kagi will be great out of the box. But there are also people who want different options. They want to be in control of the search experience, and Google and the other search engines do not provide them with the level of control that we do.
Everybody has different preferences and our goal is to provide tools, and empower users to get the experience they want, versus us trying to be smart, and creating an average search engine for an average user which doesn't exist. I mean, there's no average human, you and I are different, and everybody is different, and we like and prefer different stuff.
Nobody has their home screen on iPhone looking exactly the same. I think people deserve that in search, especially being such an important part of our lives. We probably spend half our day in web search, whether for work or private needs. It's one of the most used products out there, and it's astonishing how little customization you can do.
So after great search quality, great speed, and unmatched privacy features guaranteed by our business model, customization would be something many users appreciate about Kagi.
DKB: How do you think about ranking results?
It's pretty complex, but basically we try to figure out what the query is about, and then prioritize results from sources we think will give the best quality of results for that query. I'm speaking in general because it's a general topic.
We do have one bias actually, that we are open with, and that is we flat-out “downrank” sites that have a lot of ads and trackers, and we promote sites that have none or very few. So you could say that we have an anti-adtech bias, if you would, but unlike other biases I think this one is actually useful for the user. I think it enhances the quality of search results, which the search engines should be all about.
In addition to that, the user then has the ability to downrank or even completely remove sites they don’t want to see. These settings are currently per user. We are not crowdsourcing it, although it's an interesting idea to get all the sites that users, on average, have banned, and try to find a pattern there, and maybe we can enhance the penalty for those sites, so that new users enjoy the benefit of others basically curating the results for them. So that’s the end goal for that, but we’re still not there.
DKB: I want to talk about advertising on the web.
I talked to someone at Google about this and their response was something like “Oh, you’re going to penalize people just because they’re trying to make money? These are just normal people writing their food recipes, and they have ads on their site because they need to make a living.”
What are your thoughts on this, and online advertising in general?
Vladimir:
I think the construct of “ads are needed to help creators” is a default answer promoted by adtech to make it sound as if we cannot live without it. But the truth is that ads are by far the least profitable way to monetize quality work, because you need an astonishing number of impressions to earn money.
In reality, what this creates is an environment where very few people take most of the ad revenue, whether these are YouTubers with millions of views, or big sites, and actually the vast majority of creators, small bloggers, and YouTubers make pennies, if anything.
So most creators are actually worse off with this arrangement of things. That's one problem, and the second is that because adtech marketing is so powerful, and we are led to believe this is the only way to make money, people don't even try alternate methods.
Kagi exists because we dare to try something else. We want to monetize our work in a way where our incentives are aligned with the incentives of our users.
Most people use ad-supported models because everybody else is doing it, and that makes the problem bigger because ad-supported business models incentivize the creation of large quantities of low quality content.
Even if you're making a few dollars a month from that, you're ultimately not contributing to the web in a good way. I think that needs to change, and our mission is to humanize the web, and stop this cycle, this pattern of behavior where people just do what everybody else does, and it kind of dehumanizes us, and who we are, and we are just a part, a gear, a cog in the huge machinery of ad-tech.
The internet is run by adtech right now, and all the biggest companies are thriving on that, and I think the sooner we get out of that, the better. It is certainly going to be hard, and I'm not against ads per-se, but I'm against ads everywhere, ads by default, ads making our lives crappier and more miserable.
Users are already beginning to rebel against this model by using ad blockers on browsers. Over 760 million devices have ad blockers installed. More and more people are willing to pay for ad-free services, like millions pay for YouTube ad-free.
The ad-driven model spawns bad user experiences, privacy invasion, and constant tracking. This is especially pernicious with young kids, who are using these services with no controls, and with no understanding that ads influence their behavior, or that their profiles are being built from a young age and will be bought and sold around the internet.
By now you can probably see that I strongly oppose ads and ad-tech, maybe you could even say “too much”. But yeah, I don’t want ads anywhere near my kids or our users.
If you want ads, there are plenty of options elsewhere, but with Kagi we will go to great lengths to make sure web search is exactly what it’s supposed to be, and nothing else.
DKB: Do you have any last words on why people should consider switching to Kagi?
Vladimir:
Try Kagi to get a glimpse of what the ad-free web search experience looks like. And if you choose to support us by becoming a customer, you will dramatically increase our chances of making a lasting impact. Thank you!