Alternatives to pay per use search for Standard/Starter plan

KagiForMe Someone mentioned paying per block of searches. For example you can pay for 200 searches for a flat rate of like $3 for example.

Stripe fees should be lower that way, which was one of the reasons it was removed.

(I believe it was putting them at losses, which is obviously not good for sustainability reasons).

Okay, first of all sorry for the little tantrum I've thrown on Discord. I had just noticed the updated pricing and all the positive PR speak about it which had really thrown me off.

I'm one of those users negatively affected by the new pricing models.
I was a very happy user of the Standard plan for the last few months, doing ~ 400-500 searches monthly and paying roughly $6 each month for that. Well, with the new pricing (or rather the omitting of pay-per-use) I'm forced to either switch to the Professional plan or leave Kagi completely as only using Kagi for most of my searches doesn't make sense at all to me. But even if I were to take advantage (10% off) of the yearly subscription (which I really do not want), I'd be paying $9 a month which is (depending on my actual usage) almost a 50% price increase. Let that sink in for a minute.

I'm really unhappy with the new changes and the way they have been communicated in the announcement.
Don't get me wrong, if I was a Professional user, I'd be happy about the inclusion of unlimted searches, but I am not and for us Standard users the experience has effectively been worsened (or made more expensive). That's not exactly a great feeling, especially when every other subscription service in existence is already continously increasing their prices (and I haven't even mentioned record inflation and cost of living in Europe).

At the new price point I also no longer can recommend Kagi to any of my peers, online or offline, stranger or friend. To be frank with the current pricing I probably would have never become a Kagi user myself in the first place.

I for one do not appreciate the new pricing model and welcome the discussion about buying search blocks. I don't know what an adequate price would be, but I'd say a block of 5.000 searches for $60 should be a fair value. Maybe think about discounted pricing for larger blocks, too.

At the current pricing unfortunately I now have to think about whether staying with or leaving Kagi, at least until they come up with a better pricing model.

Someone on Discord mentioned one could cancel and resubscribe their Starter subscription when the 300 searches are used up to receive a new contigency of 300 searches. This is essentially the aforementioned "search blocks" option (albeit a bit cumbersome to handle right now).
I'm not saying this is the solution, but it might be what people need to sit out the time until there is a better pricing model available.

EDIT: Okay, @Thomas (hopefully that's his nick here, too) asked to not do this, as they are currently crediting canceled subscriptions depending on the time left, so this effectively discounts the whole approach making it unsustainable. So, please don't do this for now.

Thanks for bringing the discussion here.

Here are a few constraints that we are working with, that you all should be aware of:

• We want to remove the pay per use component from Starter as it was the only plan having it. Removing it allows for less complexity on our backend and also the possiblity to add annual payment option for Starter plan, which is something many people asked for and is impossible to have with pay per use

• More importantly, it allows us to remove mentions of pay per use from the pricing page which was a major deterrant for new customers (we heard this a lot!).

• Average cost of search is 1.5 cents (this includes projected cost of AI features, such as quick answer, discuss, universal summarizer and the upcoming Assistant, which can vary a lot per user but this is the best estimate we currently have).

• Payment processing fees are 3%-5% + $0.30 (the fixed portion is brutal towards micropayments)

What this means?

Standard plan at $5, for someone who is making 300 searches every month is at best break even for us (from a search + payment processing cost perspective) but a money losing prospect long term because we still need to pay for salaries and operations. The only way Standard plan currently makes any money for Kagi is by law of averages - people searching less than the plan allows.

Buying a block of searches as a plan is one of the ideas suggested here. The economics for this are very different from subscriptions because a block of searches is guaranteed to be used every time to the fullest.

For this to work from an economic perspective, we would need to factor in at least 20% profit margin to pay towards salaries and operations.

So 5,000 searches would be roughly $75 (cost of search) + ~$5 (stripe fees) + ~$20 (Kagi margin) = ~$100.

So to sell a 5,000 searches block Kagi would need to charge $100 to be able to stay in business. 1,000 searches would be $20.

Then we could have a no-subscription, pay only for what you use plan, that you top off by buying blocks of searches.

Another consideration is also complexity of building this and ultimately impact, as according to our data only 2.7% of our users used the pay per use component (on standard and professional plans). Vast majority considered pay per use to have an inhibiting effect on their use of Kagi.

Hope this gives the much needed context from our end so that we can have productive discussion.

Hey @STM, this is @Thomas on Discord. Thanks from mentioning the please don't do

So I've done the math for a 300 search block as per Vlad above. It comes out to (((300*1.5)*1.2)+30)*1.05=599, call it $6.

Same math for 500, your stated use per month - (((500*1.5)*1.2)+30)*1.05=976, just shy of $10. We could take a haircut on the margin given the number is a bit larger and call it $9, potentially, but that's still more than what you paid previously.

So how come it's more? Well as Vlad said, the average Standard user did not max out their plan. Previously, you benefited from people using less than their quotas and essentially subsidising your search queries. This is sustainable as long as the average on a given plan works out, and then essentially the community redistributes its allocation based on needs. If we go full pay per search instead of pay per month, then no such averaging effect occur given you will eventually use all the searches (and thus we will incur all the related costs) regardless of other customer usage, and so we need to ensure every individual search is profitable to be sustainable.

Given that $9 is the price of unlimited if bought annually, I don't think it will make much sense to go this route unless you have a very uneven pattern of use, and then you could essentially buy a big block and use that as you like over many months. If that's a use case that you'd like covered, let us know.

Stripe fees for a $6 payment are brutal and close to $0.60 which is basically 10%. That would mean Kagi margin would be about 15%, otherwise the math checks. Not ideal but I guess we could make it.

Ideally the blocks would be bigger (such as 5,000 searches for $100 or 1,000 for $20) so that this effect is less pronounced).

Hm, this seems like a hard problem to tackle. So for the time being it seems like Kagi usage will simply get more expensive for me and my fellow Standard + pay-per-use brothers.

In the long run, do you have any options to decrease your cost? Either on the search side or on the payment processing site? Will bigger volume result in a more attractive fee structure at these payment processors? Can you somehow make better deals?

From what you've written here it seems like you're generating too much cost per search. Don't take this the wrong way, I'm not an experienced business man, but if this is your cost floor, I don't really see how the business can scale given that you want to pay your employees fair wages that ideally increase faster than inflation.
Compute, storage, traffic, etc. will probably get cheaper in the future, but I wonder if that is enough.

It's a tough situation to be in, because it is really hard for me to justify to pay $10 a month for my search engine (or to recommend it to peers for that matter).
It's not that I can't afford it (it's just $3-4 more than I pay now), but I really feel like a service like this shouldn't cost as much. And with the current pricing structure you're most likely only (or mostly) targeting progammers or other powerusers, while "regular folk" don't really fit in the provided brackets.

On Discord you mentioned there were around 200 users affected negatively by this new change. According to https://kagi.com/stats that's >2% of your user base (Edit: You've mentioned 2.7% yourself above). While that doesn't seem like much, I assume this is the bracket were most of the population will find themselves in. So if you ever plan to scale your service beyond powerusers, these people have to be fitted in somehow.

I don't really know what the solution is here, I can only tell you that I won't be able to convince any of my friends to use Kagi when they're supposed to pay $10 monthly for the privilege. Even I have a very hard time justifying it and I'm basically as degoogled and decentralized as it gets and have already enjoyed roughly three months of Kagi usage. So if you can't convince me, you certainly won't be able to convince average joe (although on average people act less economically than me, so who knows).

All I can really say is that a price hike of +50% leaves a really sour taste in my mouth and I feel like you haven't quite dialed in that pricing yet.

Edit:

Vlad Ideally the blocks would be bigger (such as 5,000 searches for $100 or 1,000 for $20) so that this effect is less pronounced).

Unfortunately even at the bigger block size that pricing isn't really compelling. 5,000 searches is what I'd consider the ceiling of what I use in 10 months, judging from my usage so far (~ 3 months). For $100 this basically boils down to $10 per month, which is even more than the annual unlimited plan. (Granted, we're talking about the ceiling here, it's quite possible I could stretch 5,000 searches over 12 or even 13 months, but that still gives me ~$7.70 which is still more than I've paid until now. To be fair, it's hard to determine my exact usage by just three months of usage so far.)

What if it were left up to a reseller to figure out? Some of their plans/tiers might turn out to be able to be free, who knows?
What might a reseller program look like or enable in this regard?

Just chiming in as another Standard + pay-per-search user who fell through the cracks here.

$5 + whatever extra searches I used was very palatable, but $10 per month feels like quite a big jump and is kinda hard to justify when I have so many other subscription services to pay for. I usually ended up paying around $6-7/month, so this is pushing me to pay more, and I have to decide if I want to re-commit, and whether the value proposition is worth it.

It's also quite hard to recommend to friends now. $5 was quite an easy recommendation for a quality service like Kagi, but for $10 it has to be REALLY worth it, and a substantial jump in quality from the competitors. Maybe I can just about justify this to myself as I'm a 'power user' (or apparently a 'starter' as I'm now classed as), but for a lot of my friends/family, I don't think they'll agree.

I don't know what the solution is. Maybe increase the number of searches on the starter plan?

Either way, thanks for the quality service. I've really enjoyed the experience up until this point.