Running costs of a new SaaS
Hey IH,
I'm designing online gallery web app, where people can upload their images, add text and create gallery and share those around. Now I will be pitching it to my brother and basically asking him to build it for me.
Let's say we build it and start promoting it. Lets say there's no signups first three months. Aproximatelly how much would the running costs for such an app be?
- Landing Page
- Web App with site builder
- Three templates built
- Each user gets his own subdomain link with sharing options
- other stuff I dont currently plan
Thanks :)
Money is the easy part. You need to factor in the emotional cost of no sign ups for 3 months and the cost on the relationship with your brother when you made him spend all that time to build something that sat with no sign ups for months. You might think I'm kidding but it's true...
Around $20 sounds reasonable for the money side of things though.
Thanks. Yeah, I don't plan to force anything on him, I'm totally fine with finding someone else / abandon if no-one is really interested and believes in the idea. He's an entrepreneur, working full time and building mobile apps after hours, so he won't take it unless he really does believe in it :)
True. There setup costs as well as running costs (domain name registration, SSL certificate (which will need to be a wildcard certificate of you're generating user subdomains), etc. And don't underestimate the cost of your own (and your brother's) time. (Which is a cost, even if it's not a $ cost.)
"Site builder" covers a lot of possibilities and I'd suggest being very disciplined about the "minimal" in the MVP. Minimize the features, minimize the emotional investment.
You can now get wildcard certificates from Let's Encrypt for free so that will reduce the setup cost: https://letsencrypt.org/2017/07/06/wildcard-certificates-coming-jan-2018.html
Not factoring in the cost of the domain, then $0 with Firebase.
Why not start promoting now and build when you're sure it's worthwhile?
Largely agree with @Wouter. In fact, stick it on Google App Engine (or similar), with no customers you won't go anywhere near the free tier limits. No need to add CDN yet. So the minimum cost with zero users could be as low as $0.
However, that's not likely to be the cheapest option once you have lots of users, lots of images (all in stupidly high resolution). Think about how many photos a user is likely to have, and then think about how much space that needs. Remember that a family of four goes on holiday for a week, with four mobile phones, comes back with hundreds of photos and they'll do that every year. And as phone camera reolutions increase, so do file sizes. Are you going to allow unlimited resolution, or will you optimize images when they're uploaded?
One problem with this kind of app is that the (storage) cost per month for a given user will steadily increase, whereas your revenue probably doesn't (unless you have some kind of usage-metered billing, fair-use policy, or some such).
Then you can start pricing storage options. The best option might turn out to be one that has no free tier and includes a monthly fee even when you're storing nothing, in which case that will impose some cost even before you have users. Your brother's technical skillset will also have a bearing on this. Ultimately, it's the choice of technology and hosting service that will determine how usage (or lack of) translates into cost.
Google cloud is a resource suck. That, or we didn’t optimize it properly.
We burned through the free credits in three months while in development mode.
I've certainly seen this complaint about GAE before. You do have to look at its pricing carefully and design your app accordingly. Their pricing model is designed to encourage you design your app to work the way that suits them. Or rather, to punish you if you don't.
In particular, the pricing for storage tends to be relatively high per-access, especially per-update. The reason for this is that Google's storage technology is far more efficient at storing large chunks of data, and not at storing small chunks. If you try to use it in the highly normalized way we were all taught to use a relational database, it will bleed your wallet. If you store data in small chunks, of course you need a large number of reads and writes. You have to use it in a more CouchDB-like way, storing an entire object tree in a single chunk. Also, very careful use of indexes is necessary, as a separate update is requried for each index as well as the data itself, meaning lots of indexes results in huge costs.
Ah, I must’ve missed that fine print.
As for the pricing - it’s difficult to understand to say the least. Maybe that was done on purpose?
Or you may have used a different product than I did - I'm going back a few years.
Yes, I agree, I find a lot of app hosting platforms have confusing pricing: a fee for reading, a fee for writing, a fee per hour of CPU, per GB of data transfered, per GB of data stored (possibly with different rates for "fast" and "slow" storage), ... It can be very hard even to guess at how much some services are going to cost you before you've built your product, and even after (by which time you may well be locked in to some extent).
I used App Engine for a few small services a few years ago, mainly things that glued other systems together, where reliability was important, and the volumes of data were not huge (though the flow of data was fairly continuous during the working day). The reliability was excellent, the development easy and the need for management pretty much non-existant, and with the volumes we had, we never exceeded the allowances of the free tier.
It's not what I use for hosting my product - that's on Digital Ocean servers (a bunch of them, at $5/month each). And GAE's pricing model was a big factor in not choosing GAE - not because it's expensive, I just had no idea how much it could end up costing, and whether it was expensive or not!
Thank you very much and it's a valid question. I think I have the pricing figured out, so storage shouldn't be the biggest problem, but will test it out thoroughly.
If there are no signups the running costs can be kept to a minimum. With a cheap DigitalOcean server and some S3 + CDN costs you would be done for 10-20 dollar a month.
Unless you go for managed services like WPEngine for a Wordpress website, Heroku for the back-end, etc. Then you would be looking at something like 50-150 dollar a month depending on your requirements.
I would be more interested myself in what the running costs are going to be down the line. For my own projects i always create an excel sheet with some estimations of costs per user/client and then estimates for the first month/half-year/year/etc linked against revenue.
Such estimates are especially important in projects where resources are in a direction correlation with the amount of users (server rental, image hosting, etc). It could be part of your decision if it is going to be a viable project or not.
Thank you very much.
This comment was deleted 8 years ago