Huh, $10 a month is pretty steep. It's great that they offer a free plan, but that comes with all the misaligned incentives again. Any reason they don't just do pay-per-use (1 cent per query)?
In our survey 90% of users told us they preffered fixed fee over pay per use and feedback we got was that pay per use would make them anxious to use search. Also it adds additional friction in the signup flow (where the idea of paid search is already a novelty and then pay per search?) and so we decided to go with a fixed monthly price.
Sweet spot would be $15-$20/month but this way we would not have enough users, and less users equals leas feedback to build product. Our pricing is subject to a change, we had to launch with a price and we’ve chosen one that was good compromise.
We are likely to introduce pay per use first in our enterprise plan. Pricing Kagi is an extremely difficult intelectual challenege. (Kagi founder here)
I would bet that my usage most months would not exceed $5, and I think many other prospective users would be willing to gamble on that as well. Food for thought.
* Say 1 search is 5 cents. So unless a person searched less than 200 times a month, the pay-per-search is cheaper....
* Or why not adopt the model where you pay 5 cents per search _for the first 200 searched_ and nothing afterwards (so the ceiling is $10 per month)?
Relevant Tweet: https://twitter.com/janhodl3/status/1535530318987509761
At the moment, I'm either overpaying (because I perform less than $10 worth of queries per month), or the company is losing money on me. And with the existence of the free tier, the business model can only work if most paying users are effectively overcharged significantly. Right now they are operating at a loss in both tiers, if their pricing page is to be believed.
One would hope that costs amortize better with more users (e.g. scraping is pretty much fixed cost regardless of the number of users, but maybe that's already negligible) to push the price low enough for pay-per-use to not feel spendy. (When did you last think about how much one toilet flush costs you?)
The company is aiming to have many users that everyone is overpaying with their $10, so that they make money, thanks to the reduced marginal cost. And the company hopes that the $10 is low enough that enough everyone knows they are overpaying they are still willing.
Correction: they're saying 80 searches cost them ~1$.
>Why does Kagi cost $10/month?
>Our proposed price is dictated by the fact that search has a non zero cost. With other search engines, advertisers cover this cost. But it costs us about $1 to process 80 searches.
>Someone searching 8 times a day would perform about 240 searches a month, costing us $3. An average Kagi beta user is actually searching about 30 times a day. At $10/month, the price does not even cover our cost for average use, and we are basically betting that average use will go down a bit with time because during beta people may be searching more than normal due to testing etc.
>Our goal is to find the minimum price at which we can sustain the business. If it turns out that we have more room we will decrease it. But it can also be that we may need to increase it.
>The free plan will be limited to 50 searches a month (and this too has to be paid by paying customers which makes the above math even harder).
8,000 searches costing a dollar, in actual resource use? OK, maybe. Still seems a little high, but maybe. 80? Are they paying someone to manually look things up for you?
You don’t have to answer, but how do you have such a high per search price?
I'd challenge that the cost per search is actually very low. For just 1.25 cents, you can search the entire world wide web with any query, find what you need with no distractions and all that in just 500ms. The value of that search outcome for you can be $10 or even $100. It is probably the most amazing return on investment existing in the world today. Imagine for a second that Google and other search engine didn't exist, and someone offered you this deal today. I'd be you'd take it in a hearbeat. Heck you may even consider paying $1 per search or even $5. So the value is there.
But today this value is anchored in the price of 'free' search. So instead $1 per search sounding like a good deal, we struggle to accept that 1.25 cents is.
We hope that for a portion of people that 1.25 cents per search is going to sound like an amazing deal, especially as this is the only search relationship where the incentives of the information provider are aligned with the incentives of the information consumer.
I don't feel 1 cents per search is a good deal at all, while -logically- it probably will give me more than 1 cent of value out of it. People never thought about how much the knowledge they acquired is worth, most of the time they share it free of charge, people like to share, somes will even pay hosting to share knowledge !
That's why I don't believe in the pay-as-you-go I'm fine with a subscription, I don't pay to have more value out of my search, but because: 1. I dont like ads, I want to support an ad free alternative.
2. I pay for my privacy.
3. Better results will mean less time lost on searching, less frustration.
I never accounted the value I will get out of better search, I looked at the price and felt that I was a bit overpriced for my taste, but I want an ad-free, pro privacy internet.
Which is probably why search quality is going down. They're rewriting your query to a more common way of saying the same thing, at least according to Google.
Date-search count-generated cost
>Jun 10, 2022 90 $1.12
They show conflicting numbers.
Sure, but this whole adventure won't last very long if the company loses money even on paying customers. If your usage costs them about $30 a month but you only pay $10, who will pay the remaining $20? Someone has to finance your access to knowledge in the end...
How much does it cost to have a suboptimal search experience 50-200 times every day? Saving 5 seconds (on average) per search, that's something like $10 per day in savings (provided you search during work hours).
Nevermind. Looks like you have to have a subscription to see usage stats.
As much as it bothers me, I'd prefer to work around the first option.
I understand they need it for customization and monetization, but search queries are too private to ask for trust.