There's a few areas of app development that we know are unnecessarily difficult for developers. Ideation, branding, marketing, and pricing for example, are all topics that could fall outside of our typical skill sets.

Let's try to articulate even further what the difficult components are, and see if we can come up with some ways to alleviate developer pain.

Let’s talk about ideation first. How did you folks come up with the ideas for your Shopify apps?

Laziness. We were doing something manually, and so we wanted to automate it.

It definitely helps if the idea comes from experience using Shopify, or from a lot of exposure to clients that are using it.

Have you gone on to build any other apps beyond the one you build to meet your needs as a merchant? We've been finding that there isn't enough communication between merchants and developers to get the ideas flowing directly. I'm wondering if there's a better way to make those connections than sending people to the forums.

My first app idea came from Edward. :)
The majority of ideas come from my Shopify clients.

Do you think that having a client base is vital to app ideation without some kind of Shopify intervention?

No, I don't think so. Coming up with ideas is easy. Execution is the hard part. If you don't fully understand the problem, how can you solve it? Having a client base at that point, or being a merchant yourself, is vital to having a great product.

Pricing web apps is hard; it was the number one ask for help from Shopify App Devs in a recent survey. How did you go about setting your current price?

That was one of the hardest things for me. When I first entered the App Store, the pricing was all over the place. It really depends on the type of app and the value it brings. I've experimented with pricing to see what people are confortable paying. I first launched with low prices, but eventually increased to where it is now. I think I've hit the sweet spot for what my app does.

To me apps fall into one of two categories: Feature or Add-on. A feature is something simple that could be in Shopify core (a contact form app for example). I don't think merchants would be willing to pay more than $5/month a feature app. An add-on is more complex (cart abandonment, recommendation engine, etc). Merchants are willing to pay $100s if it adds value.

Do you think that one time charges for apps would work better for developers? We've been seeing some good results with initial experiments. Do you feel like the cost of maintaining a customer's data and potentially providing support over time is insignificant enough?

Depends on the app. For Jilt, there's no way I could go to a one-time fee because of the hosting costs.

I would love to see Shopify develop a plugin framework that allowed merchants to upload plugins to their shop (similar to themes), but I understand the security risks with doing that. If there are no other costs involved besides support, then I'd definitely do a one time fee for certain apps.

Yan: are the hosting costs really more than $60 a month? I bet you could set a one-time fee of $80 and see the hosting costs recouped as long as you made at least one sale per month.

Given that Shopify is growing like crazy, I’d say you’d be way more than just safe with a pricing strategy like that.

@edward about the idea: (Sorry I can't branch from Branch... I hate beta apps that limit people's ability by waitlist...)

I came up with the idea for DeepMine because I was constantly downloading XMLs and running filters over them. I use the example of "number of unshipped international yo-yos" because that was the tipping point example, I got so damn tired of having to look up each query each time for something that could just be generated. Since at the same time MarketBrain was doing some kind of weird transition and just straight-up wasn't working, I took some time and built DeepMine.

(ugh Branch has a character limit?)

When it comes to pricing, I just used the $10/month that MarketBrain charged because it seemed a fair price to many. Too many apps are charging WAY TOO MUCH and as a shop owner who can program, I'm like "screw this, I'll just make it myself".

Example: I really hate ExpressRex when they upped the price from $10/month to $70/month for unlimited plan holders. I'm still VERY grumpy about that.

When it comes to one-time prices, there's no way I feel comfortable doing that. Server pricing is way above what I could make with one-time pricing, and it's much easier to get people inside at a nominal fee instead of a one-time $100 fee.

(char limit)

Good job on heading this up Lyd!!

This is my first Branch post, I'm new to this. I must be spending too much time working on apps I had never even heard of this Branch thing, Ha! Looks like a good place for Shopify Discussion, I like it.

Jason: Glad you like it! We’re testing it out as an alternative to the forums in the hopes that we can drive more active/useful/regular public conversation here.

Ryan: regarding pricing, we’ve found that via a first pricing experiment, customers are more willing to go with one-time fees than we thought. We’re also really jealous of how much money [Magento Extensions](magentocommerce.com) are making, and it’s all one-time pricing there.

I'm with Yan, in that if each install generates more load on the servers, and more customer service demand, then the only way to charge is with a recurring charge.

When there are as many Shopify stores as Magento stores, it will be a different story. It would be nice to do a one-time charge.

With regard to ideation, I think that connecting to services is a pretty good place to start. While only a small part of the overall app eco-system, it would be good to get a running list of the external services (Mailchimp, Salesforce, etc) that merchants want to be able to integrate with.

We've been testing the one-time fee with the Product Discount app have seen good results so far. In fact in the first two weeks we generated more revenue than the previous 2 months. The feedback from users has been mainly positive as well.

It's still a bit scary though as there's no good way of offering a free trials and a number of other things but that could be worked out eventually.

I have a whole list of pros/cons, ideas and feedback we've got from users that I'll be sending after testing it for a month. We've done two weeks now and it's been very interesting.

We're also doing some profit analysis. Right now our One-Time fee is the same as what 5 Months of the recurring fee would have been before.

Jason: awesome to hear! Your experiment is definitely going to be a case study coming up.

What’s frustrating as the guy watching over the App Store is that I know that app publishers could be a lot more successful, not only in terms of customer base growth, but also revenue-wise.

As part of an upcoming revamp of the partners dashboard (can’t wait to show you guys some mockups of what’s coming…), I’ve been talking to theme designers who have played with their pricing has been eye opening – I’ve heard lots of cases where lower pricing actually hurt sales in comparison.

When Shopify started the Theme Store, we knew that we wanted to offer high-quality, premium themes, and that having a $100 minimum was going to be part of that offering. We knew if we had no minimum, that it was going to be a race to the bottom, and quality would suffer.

In retrospect, I wish we had a minimum price as well.

Brian: we’re at 30000 shops so far, and 56% are aggressively using apps. Shopify’s growth rate is still looking like a mean hockey stick, so I really do think it’s worth trying out a one-time fee pricing model, even if it’s an optional $50 fee.

I’ve got to conduct more interviews with merchants to confirm this (as well as watching what happens to the few apps switching to one-time fees), but I have a feeling that purchasing apps straight-out is more comfortable for many merchants. Heck, most apps are going to cost less than a video game or dinner for two.

But one-time fees do nothing to solve my Heroku bill, which is quite large. I have to get out there and hustle to try to get new sales all the time from other sources.

Whereas with recurring billing, I can rest easy knowing that I get paid a steady amount per month, and I can easily afford Heroku with proper budgeting. If I need to double the Postgres cache (again...), then I safely do so.

And Magento is on a different scale: Most Magento apps pay for the software, and use their own machines and databases to run the apps. With me, I pay for the Heroku databases.

Plus Magento extensions have the ability to query their database directly, whereas with Shopify I have to download all orders to do what I need.

With Chimpified, our first pricing structure was based on features. To get at the better features, you had to upgrade your monthly plan. However, after a year or two it was clear that was a bad way to do things.

The app works by responding to orders, so the more orders that a store generates the more it should be charged. We found that the stores causing the most load were on our free plan, and the ones that had upgraded weren't generating much at all. So we decided to add monthly order limits to our plans, so that we would be charging for what was costing us money.

So far, so good. We've seen an increase in the number of people upgrading, and there doesn't seem to be any blowback.

Edward: Agreed. I think that the merchants would prefer a one-time fee too.

You know, the Magento thing does raise an interesting idea. What if Shopify could internalize the hosting of the apps? App developers create apps, deploy them to the Shopify hosting solution, etc. With no hosting costs to worry about, and scaling not being a consideration, a one-time fee for apps would be similar to the one-time fee for themes.

I'd be much more comfortable with one-time fees if I didn't have such a large database bill.

Are there not ways to host your apps less expensively? It surprises me that you guys are having to pay so much. If you absolutely wouldn't consider one-time pricing, is there anything that would have made establishing your pricing plans easier?

Ask HN: [Pricing] Charging a one-time fee for a SaaS app?
news.ycombinator.com

Fielding webhooks (in our case) triggers background tasks, and with thousands of stores generating thousands of orders each, hosting doesn't stay cheap. I think we're running 8 background workers at the moment.

On the hosting costs: yes, there are ways that are *much* cheaper than Heroku. Hosting with Heroku is insanely expensive. The Linode VPS that hosts our app costs $80/month, whereas the number of app / worker processes that we run on there would cost $481/month on Heroku.

Right now Shopify encourages Heroku usage for apps, which is probably fine for a proofs of concept, but it might make sense for Shopify to also provide e.g. a Linode stack script (or the equivalent for some other hosting platform).

The real ongoing cost of accounts is support. If you work out the math, it costs about $9-15, depending on salaries, to handle a single support request. Once you work out Shopify's cut, that means for every support request that comes in, you need to have billed at least $11.25-18.75. (That's assuming $60-100k salaries, 1.5x salary overhead for an employee's cost and 10 minutes to handle a query -- some are much shorter, some much, much longer.)

On a one-time pricing scheme, that means for a $100 app, you'd need to make sure that you average less than 5-9 support requests over the entire customer lifetime just to hit break-even.

Scott: I spoke with some Shopify customers who use apps last night at an event we held about how they felt for paying for support and they were all game, even if it’s something like $10-$15 per issue after the first few.