Running Costs of a SaaS app | Hacker News Reader

Running Costs of a SaaS app

cushionapp.com

278 pointsjanpio115 comments

Worth noting: costs for people blow costs for software/services out of the water. I was notoriously spendthrift with regards to SaaS services ("gotta catch them all!") and spent ~$3 on people last year for every $1 on services. (And if I had had any FTEs that would have been substantially higher.)

Over ~10 years running SaaS apps I generally managed to pretty consistently find ways to spend 40% of revenue.

Also worth noting: as you get further from "things which are made by devs-running-businesses-as-hobbies-for-devs-running-businesses-as-hobbies" towards "business inputs sold by businesses to businesses" costs skyrocket. You're going to spend plural thousands of dollars on bookkeeping and tax compliance every year. You will eventually need a contract reviewed; that will probably cost you a few hundred bucks. You're going to get a business insurance policy; you'll find the bidding starts at a grand. etc, etc

I found it frustrating (but also understandable) that the running costs apparently don't include any personnel when they almost certainly dominate the entire expenses.

:waves: Hi, I'm the founder of Cushion. I'd be happy to answer any questions about our running costs or transparency. (Thanks for posting, @Sujan!)

Ha, that's what you get for finally posting to the journal (How it's made blog from Cushion [1]) again. That's why I still love my feed/RSS reader.

[1] https://cushionapp.com/journal/

Heh. I was curious if anyone was still subscribed.

Hi, I was interested in what the crossed out lines represent?

Hi! The crossed-out costs are ones we no longer use. We either no longer needed that service, or we found a better alternative.

Do you regret spending $2000+ on a domain name?

Ha! Not at all. It was a big expense, but I think it’s worth it for how we use it. The domain is "in.vc" and it’s used when sending an invoice link to clients. Instead of sending a super long link and possibly confusing clients with our app name, they see a short, generic link.

I wish we had cushion.com, but last I checked, it sold for $120k in 2012, so that's off the table. :)

Is Cushion trademarked?

Yes! I should actually add that expense, too... Thanks for the reminder. :)

I was actually asking because I got an ethically-questionable, evil idea: If you are owning the Cushion word, can you sue for the domain name cushion.com?

No. The trademark must be registered and exist at the time the domain was registered and that the domain owner knowingly and intentionally registered the trademarked domain name.

how sure are you of this? I have owned a domain since 2003. a few years back a new startup named <domain> inc tried to buy it from me. they ended up getting domain.org but I checked recently and they registered the trademark. I was reading that I have only a few years to appeal it showing I am harmed by them. I did not read anywhere that they could not sue me

Not legal advice, but fun reading: http://nissan.com/

If you mean trademark, it must be registered under one or more of the many[1] categories of business in which a trademark can be registered, which means the same mark can be registered by different companies as long as there's no "substantial likelihood of confusion," so whatever you mean by "owning the Cushion word," it's going to have to be something else. It's not as easy as registering in all possible classes, which in itself is quite expensive.

1. http://www.nolo.com/legal-encyclopedia/trademark-classes.htm...

Any particular reason you moved away from Mailchimp? Do you do any form of email marketing now?

We switched to Intercom for sending our monthly email simply because its email editor is so good. I wish a proper email marketing service would come close to their simplicity, but I haven't seen anything yet. We might be moving back to Mailchimp soon because the per-user cost for Intercom has really skyrocketed since we've used them for email marketing. Granted, Intercom actually don't want people to use them for sending batch emails. I wish they'd come out with a standalone service for it—remain as dead-simple as their editor is now and they'd make a killing.

Have you considered Drift as an alternative to Intercom? I have a small SaaS app as well, and my Intercom bill is about $130 a month compared to $50 with Drift.

I noticed them a while back, but I'm happy with Intercom. They're good people.

Great stuff!destroytoday!

Regarding Heap Analytics, their pricing model is quite confusing. I'd like to know how many users/visitors to your website are you getting that allows it to be free.

To you, is the almost $5k you're spending on Intercom worth it? How much does it come to per user?

You moved away from Baremetrics to Chartmogul, something with not as good a design. Why?

I'm also curious as to why you moved from Baremetrics to ChartMogul.

Thanks!

In all honesty, we're way over Heap's free limit, but they haven't throttled us or reached out, so shhhhhh.

Intercom is so worth it. We wouldn't be in the same position without it. The reason it's so high (as I mentioned in another answer) is that we also use it for sending our monthly newsletter. We could definitely get the cost much lower, and we're planning to, but I just need to sit down and do it. We've got our hands full with our current roadmap, so I haven't had the chance to move everyone to a more affordable email marketing service, but sooner than later.

When I used Baremetrics, they were changing their pricing a lot and experimenting so much with growth tactics that I don't agree with, like removing the ability to cancel on your own—you now need to contact them.

I tried ChartMogul and they've been absolutely wonderful. The team is so kind, helpful, and responsive. They also don't make drastic changes, which is refreshing since I only need them to track my Stripe metrics.

Re: Newsletter.

I cam across Email Octopus and I found setting them up easy. You may want to check it out. https://emailoctopus.com/

You should save as much money as you can.

Any idea the new cost if you stop using Intercom for newsletters?

Thanks! Not entirely sure, but I think we’d at least cut the cost in half. Aside from user count, we still use a lot of their packages, which have a decent base price.

I run EmailOctopus - happy to discuss anything we can do to help you out, as we're a relatively small SaaS business ourself and know the struggles.

"The team is so kind, helpful, and responsive."

Founder of ChartMogul here...thank you for the nice comments, you made our success team's day! <3

Just to second that - Nick and the team at ChartMogul have been awesome to work with. The product has evolved in leaps and bounds over the last couple of years.

We tried Baremetrics originally but when we dug into the nitty gritty of the numbers, there was a lot missing. To be fair, this was 2 years ago so I'm sure it's changed a lot.

Baremetrics has been notoriously slow for the past 2 years - it should not take 10 seconds to load up static graphs (cache them and recalculate them on the backend every x hrs)

i know they have a small team but their inexperienced tech/backend team has costed them dearly (again i know they probably have worked hard but you can tell them they are inexperienced)

Heyo! Founder of Baremetrics here.

Wanted to touch on a couple of things in this sub-thread on Baremetrics.

It means a lot that literally _anyone_ has ever used Baremetrics. I started it over 3 years ago to scratch my own itch and hot dog it kinda took off. So, I'm grateful...certainly for the nearly 700 happy paying customers, but also for the others over the years who's experiences were less than stellar.

We've had some growing pains, but the majority of the painful ones are far in the past.

I did want to address the specific things folks brought up here, just to clear the air.

* Regarding changing our pricing "a lot"...we've changed pricing just once in 3 years. And that was based directly on customer feedback.

* Regarding us being "notoriously slow for the past 2 years"... We had some temporary scaling problems over a year ago, but that's just not the case any more. We've got a whole suite of server monitoring tools that show we have, in fact, not been "notoriously slow for 2 years".

* Regarding us being "inexperienced"...busted! I've never built a complex analytics platform storing and processing in near real-time terabytes of data! ;) Every single entrepreneur is winging it. You'll make some good decisions and some bad ones. That's life. :) But to be clear our team is very equipped and extremely talented. We're just solving a complicated problem...much more complicated than anyone who's never built this sort of thing before realizes.

At the end of the day, though, there are obviously some people who haven't had the experience they wanted with Baremetrics. I hate that that's the case, but we're constantly improving and innovating...most of the time for the better.

If anyone has had a less-than-positive experience with Baremetrics, please email me. I'd love to chat: josh@baremetrics.com

Thanks again for all the feedback!

Thanks for your response. I'm not a user of your service, but I've always appreciated your transparency with how you've run (and grown) your business.

Have you ever considered any Heroku alternatives? I've had quite disappointing cost experiences last time i used it.

We've thought about it, but Cushion is incredibly lightweight and doesn't need to be used constantly, so our load is very low. We've only needed 2 dynos since we launched and we're nowhere close to needing more than that. As much as I'd like to lower the cost, they truly take so much weight off our shoulders. I gladly pay what we're paying now, so it's not an issue for us.

Have you looked into Appengine Flex. Appengine flex is a PaaS like heroku, and will scale up any docker container. Its a great solution for your early days, and when you do need to move to a 1000+ node cluster its easy to switch from Appengine Flex to kubernetes. You also might get a lot of free startup credits https://cloud.google.com/developers/startups/

(Disclaimer, I work for Google).

[1] https://cloud.google.com/appengine/docs/flexible/

You should mention Python 3 support specifically.

I haven't, but I'll check it out. Seems like it's still beta? (says not to use in production)

So I am wondering about heroku: you say that it's quite expensive, but $25/mo doesn't seem that much. How many users can one $25/mo server actually power?

I think I wrote that when they had their older pricing, or I didn’t realize their real value. Also, I’m pretty sure we could run Cushion on a single dyno, but I like having two around to balance.

Aha, makes sense. Thanks!

Why is directory listing turned on

We just made the page static to handle the HN traffic, so there were a few seconds between the switch. pulls collar Didn’t expect to get featured today.

Hey Jonnie! I'm super curious about the move away from Siteleaf (esp since you're so close to the Oak guys), and how you+marketing team ended up on Craft?

Hi! Great question. I love static websites, but we hit a point where we were updating the site constantly and needed dynamic content, like the running costs (which pulls its data from a google spreadsheet). There are certainly ways to do this with a static site, but with the size of the site and frequency that we update it, I really wanted to be able to hit save and it was live, rather than wait for everything to compile and publish. I still use Siteleaf for all of my other websites, but cushionapp.com is a different beast.

Also, lol at "marketing team". I am marketing team. :)

Ah okay that makes a ton of sense with how dynamic the site now is. Looking back now, would you have still gone with a static site generator or started with something dynamic (question we're pondering now with monograph.io which is middleman)?

Also on the marketing team, I got head-faked by the growth freelancer, and assumed you went with a dynamic CMS to get more help. But man impressed with how much you still do!

Thanks for the insight this provides. Any chance you could share this spreadsheet template?

Why don't you guys use the new Heroku beta SSL feature so you can bank that extra $20 a month? It's not much but it makes a difference if you're spinning multiple sites up.

https://blog.heroku.com/announcing_heroku_free_ssl_beta_and_...

FYI I don't think it's beta anymore :)

We would, but unfortunately, we need their "old" SSL add-on to get SSL to work with the root URL of our invoice url, in.vc. Apparently, the new SSL doesn’t work the same.

There's a guide for converting your certs, it's just a format issue most of the time.

Hi fellow NY SaaS developer :) I ran a similar report for our year end books, and it's amazing the amount of SaaS it takes to run a SaaS. Or put another way, the incredible ecosystem now available that allows you to focus on your core business.

I don't see any third party Stripe dunning service, did you build something similar yourself? Otherwise it might be worth looking into, it's one of those things that pays itself back in no time.

For anyone else looking into recurring billing I'd highly recommend Chargebee (not affiliated at all).

We've tried a few different solutions and were a little apprehensive at first because it looked a bit clunky but it's actually one of the best bits of software I've used. You don't realise the number of features a subscription billing platform needs until you use something like chargebee and discover how wonderful it is to have someone do it all for you. What's really great is before changing anything they give you a pop up showing what's going to be billed now, what will happen in the future etc. We used stripe before switching and I was a nervous wreck before ever pushing any buttons.

My co-founder and I were saying just the other day that we think it's one of the best investments we've made as a company.

https://www.chargebee.com

Oh wow, this is exactly what I was looking for. Thanks for sharing!

Glad to have helped! I discovered them when trying to find a solution for the insane EU Vat regulations (which they handle gracefully).

We've been with them through a big few tasks now (like when they added Xero integration - which is also an amazing feature) and if there's any confusion with mapping data or whatever they'll always jump on a video call to make sure everyone is 100% clear on what's going on.

Care to share your revenue (or at least a ballpark?)

Thank you for the info as it helps me formulate how to do a small non unicorn start-up in a smaller fashion..very helpfull

Thus far legal has been one my most significant cost (incorporation + NY qualification). Perhaps I'm doing it wrong -_-. What do your numbers there look like?

Fortunately, I was able to find a really great lawyer who is actually affordable (previously interviewed a few law firms who were $700+/hr). Besides incorporating, legal costs haven't been too high. Just the occasional contract here or there, or updating our terms.

Tip: if you're reading this on your phone, switch to landscape orientation to reveal a bunch of detail that's otherwise hidden. Then switch to a tablet/desktop to see even more detail. It's a shame the responsive design removes columns on smaller screens rather than finding another way to present all the information.

I wanted a "total total" and calculated it. For the curious: $17,593.10.

Obviously not all months are equal, but that's about $517.45 per month.

Look like an awesome tool!

First, thanks for boldly sharing. It's quite an accomplishment to finish anything.

You guys definitely spent a lot where it doesn't count ($2k for a cool secondary domain, $1-2k for fonts and SSL, ...), but if you've made it all back then good job forget I said anything.

appname+app.com was probably not the best idea. Domain name hacks are confusing, and you are now indirectly working for appname.com

A startup's name means everything and nothing. All that matters is that it's short, relevant and memorable. clownmaker.com will always be better than clownapp.com

Alright I'm done with the unwarranted advice. Really good job.

It's surprising to see the cost of Intercom compared to everything else on the list.

We recently reviewed how much it was costing our business, compared to the value we were getting out of it, and switched to a cheaper (and simpler) solution.

>and switched to a cheaper (and simpler) solution.

Well don't leave us hanging.

Agreed. +1 this :)

I didn't want to come across as advertising something. We noticed that we weren't making use of much more than email support, so we switched to GrooveHQ. Their pricing is much more reasonable, and their story and the way they interact with their customers is great. But really, there are tons of options out there.

My point is that if you're not making the most of what Intercom offers, then it's pretty damn expensive.

I switched from Freshdesk to Groove as well.. after an aborted attempt at Zendesk.. Groove is definitely my favorite; easy and lightweight.

I think this question comes up a lot and answers for this are usually a Google query away. However, Intercom does a lot of things. destroytoday even mentioned above that one primary use is email marketing, so it's probably only fair to consider what the combined costs are. In most cases, you'd need to pay for both a helpdesk + an email marketing app.

In that light, the grandparent's post might not be a fair comparison, since he/she might be only talking about customer support functionality. In that case, there are far cheaper alternatives. Even heavy weight companies like Zendesk will likely be cheaper by comparison, but you won't get the marketing part built in.

On the other hand alternatives like Drift have been mentioned above which seem more like a fair direct one-for-one comparison with Intercom, and will usually come out cheaper.

Disclaimer: I'm the founder of Reamaze, which competes in the helpdesk (not marketing) space.

For Stripe analytics, Profitwell is free. And awesome.

https://www.profitwell.com

Looks nice, but I'd never not pay for a service because you're paying somehow, or the service won't be around for long.

Well, they have a paid tier. I feel like there's room for a little optimism when superusers are funding the existence of free plans.

Oh, okay—I browsed the website and couldn't find anything about a paid tier. nod

Hola! Patrick here - Founder of PriceIntel/ProfitWell. Why ProfitWell is free is actually a longer story, but the short answer is we discovered we could monetize more effectively (and efficiently) by giving away the metrics for free (while still being the only one on the market that's 100% accurate) and then selling add-ons like Retain (kills your delinquent churn), Recognized (Revenue Recognition), etc.

Here's our pricing page (with paid tiers - you have to scroll down): https://www.profitwell.com/pricing

If you're curious around the science/data that went into this (the longer story), you can check out a talk I gave at our SaaSFest conference: https://pi.wistia.com/medias/jdywlrpowl. We used our Price Intelligently software to determine market willingness to pay/feature value before writing too much code.

As always, let me know if you have any questions. Happy to help.

Crap—I'm one of those people who didn't scroll. :( I hate when that happens. My b.

Thanks for the info!

Absolutely. That pesky fold gets me, too (even though our design team tells me the fold doesn't exist ;)).

They have a pricing consulting business that finances it (http://www.priceintelligently.com/) so the data feeds their consulting business (I'd imagine). I use Profitwell and love it.

This info is actually very useful for me as it confirms my (rational) fear of accumulating expenses. It's way too easy to just spend $10-$50 here and there and before you know it you are spending $1000 a month, while the core service is still running on a single triad of web-server/worker-server/database-server combo with no redundancy.

On the other hand it probably contains everything I could think of and it's still only ~$1000 a month, which is not a lot if you work as freelancer and run your SaaS on the side.

Spot on. It's also tougher with individual users where $10-$15/mo is the norm. B2B services have it good.

I literally use none of any such paid services. Zero. It's really not that hard. Just automate what can be automated.

Automation isn't just a choice where you snap your fingers and it's done. Automation takes time and effort and adds technical debt and support burden; sometimes these are trivial amounts, sometimes they aren't.

I use a service that sends me a bunch of fairly trivial Google Analytics data. It costs $10/month for the equivalent of fifty lines of Python.

I know I could spend maybe two hours writing a replacement for it and set it on an EC2 host and probably never worry about it again, but those two hours are still non-trivial amounts of time: two hours of my time pays for more than three years of that $10/month. Plus, the API may be brittle, there may be edge cases I haven't thought about, and I don't want to add yet another thing to monitor and worry about.

Sometimes, yeah, you can automate things and roll your own dependencies -- and sometimes the calculus works out and it's a no-brainer. But I've found that homebrewing a $XX/mo dependency is rarely the best use of my engineering time.

These are actually solid life advice for programmers: https://xkcd.com/1319/ & https://xkcd.com/1205/

> I use a service that sends me a bunch of fairly trivial Google Analytics data. It costs $10/month for the equivalent of fifty lines of Python.

Would you mind sharing which service that is?

I run S3stat, and that's a pretty fair description of the service we provide (for that price). It brings in a nice living from small businesses that can perform the basic arithmetic described above.

Specifically I was referring to Briefmetrics:

briefmetrics.com

S3Stat (who another commenter mentioned) is also a great example.

Cost benefits though. $15 a month to save a few days coding is a bargain.

It really depends on what your turnover is, where you make the most money, and what your time is worth (objectively, rather than ideally).

Is it really important to distinguish your web application with another font - looks like he spent $500 on Web fonts, which is a lot for a startup ?

A one-time fee of $500 is worth around two to three hours of good design/branding labor. I think having a distinctive, high-quality font is absolutely worth that investment, especially when you consider that the target audience of the SaaS includes designers.

so I mentioned only $500 dollar expense for the Web application and not the Marketing font that costs $300 more - so why do you need to worry about branding for web application ? I agree that marketing websites would need distinguishing fonts. Just a thought.

It’s less about distinguishing the app with another font and more about designing the app with what I had in mind—the system fonts and more affordable fonts didn’t feel right to me.

Even though I consider myself more of a developer, I went to school for design and it’s a big part of my work, so the look and feel is just as important to me as whether the app works or not. Sure, we could’ve saved $500, but I wouldn’t enjoy using the app as much, or working on it.

I also got to support two type foundries that I really respect. (JAF & Dalton Maag)

"the system fonts and more affordable fonts didn’t feel right to me"..Thank you - that makes sense

I am just using fonts from Google Web Fonts, but when we made the change we had several users reach out with how much they love the new fonts.

Since then, I am very intentional about using fonts consistently and intentionally. I could definitely see us "upgrading" our fonts when we do a design refresh.

The website looks wonderful, and considering just how much text you need to use to get ideas across, the font money seems well spent. The entire page looks... delicious.

Thanks!

Why not switch to "let's encrypt" for SSL certificates. It's free !

I myself am curious as to why you're not using let's encrypt.

Read the whole article... they have started to use it and will migrate more of the app there... "

We use Let’s Encrypt for the SSL certificates for both the marketing website and in.vc. Eventually, we’ll migrate our other SSL certificates to Let’s Encrypt. It’s free and easy. "

Read entire article... multiple times. kind of arrogant of you to assume otherwise. no mention on the post what so ever about let's encrypt.

The issue is a responsive design that hides most of the content on the phone.

Thanks for pointing that out... in a friendly and constructure manner.

Cheers!

Didn't mean to be arrogant. Sorry for that. Had no idea it was a responsiveness issue, was reading from desktop. I'm not usually so short in my responses. Again, sorry about that.

Ahh my mistake, as another use pointed out some of the content is omitted on mobile devices.

Let’s Encrypt wasn't around when we started, but we do use them for our marketing site and will migrate the app when we need to renew our current cert.

The cost of SaaS apps can really start to add up. What's interesting is that there's a very similar "stack" that almost all SaaS apps need to buy from other SaaS apps, and many of them don't have a free alternative.

There exists a whole class of apps/services that could gain a lot of traction by providing a freemium model. By becoming one of the best solutions AND having a free version (for limited use), I think you have a better shot of becoming part of a new bootstrapped SaaS business' "default stack".

This is why we gave HelpSite.io [1] a free version that includes features that I consider "necessary" for any business no matter how small, even when most of our competitors are charging for some of them. The more you can align with a startup's own business model (e.g. they need lots of light services up front even when they have no revenue; later they will have revenue and use those services more heavily) the better off you will do.

1. https://helpsite.io

What do you guys actually do?

Cushion is a tool for freelancers to manage their schedule, forecast their income, and track invoices. Time tracking is also dropping soon.

You should check out https://wakatime.com. The API could be used for your time tracking.

Oooo nice! Thanks for the link.

Didnt see any pricing info for your voice infrastructure. Was it a conscious choice to not go for a business phone system?

So overall 17,593.1 USD for 3 year of maintenance? Nice job! What are actual cost of hosting cushion, GitHub Pages can seriously host this whole thing?

Am I reading it wrong or are you just using/paying for Slack for a single user?

Currently, we pay for a single Slack user (me), but each person who works on Cushion-related things are single-channel guests. I work with one other full-time dev, so he's in the #dev channel. I work with an editor in the #content channel. Growth guy in the #growth channel, etc. It's actually refreshing not having everyone in one channel—a lot less sifting.

Definitely not lean. From the first look - too many tools and expenses for unknown no name startup. I'm wondering how much of these can be painlessly cut off?

The name is Cushion.

“Unknown, no-name startup” seems a bit unfair. It looks like Cushion has a solid product & paying customers.

A bit harsh. I guess they are very profitable and those tools probably save them money and time.

If fame is your measure for success, sure.

Cushion took no funding to reach profitability, and is growing at a rate that Jonnie seems to enjoy.

Here's his portfolio site, in case you wonder what no-name unknown companies worked with him before:

http://destroytoday.com/

Can you define 'painlessly'? I'd also read your detailed opinion / chart if you think similar results can be achieved with alternatives - especially because OP is being 100% transparent on what & why they need to spend.

Reply on news.ycombinator.com