Launched my API and shut down my Shopify app

Long story short… my product is now a pure API. Send data, get back an image! You can test out the API in this live demo:


Long story long… this is the story of another pivot. (But a “soft” pivot this time!)

January was an interesting month. I launched a new version of my open graph image generation product and took on a whole bunch of new users and a dozen or so paying customers!

January ARR: $5486
March ARR: $8224

I also began exploring the Shopify App Store as a distribution channel. I had heard a lot of good things about it, how it gave you access to huge amounts of customers, and how the conversion rate can be quite high because it’s essentially a “one click” process (billing is done via a customer’s Shopify account).

I don’t want to say my experience with Shopify was a disaster, but it was pretty close. The silver lining is that I learned from the experience quickly and have now pivoted away into much more exciting territory.

While it’s true that Shopify gives you access to loads of customers, there are some caveats:

  • Very price sensitive, in my (limited) experience it seemed the expectation for a lot of Shopify users was that the app should cost around $5 per month. The store is filled with apps charging less than $10 per month.

  • The Shopify API is clunky. My app required integration into a user’s <head> and there’s no way to do that other than to literally use the API to massage a user’s template code directly. This isn’t very clean, and obviously can have unwanted results. A better way to do it would be to have some kind of callback / hook system, like the Wordpress wp_head API. Messy integrations leads to more customer support requests, which leads to…

  • Unsustainable levels of customer support required, at the price level expected. I was having multi-day email threads with customers, when my Shopify pricing was $5 a month. Basically losing money on each sale.

I think Shopify can work for some indie devs, but not for me. It’s best suited when you have a really zero-setup required product.

I will note that with zero marketing, on the Shopify app store I was getting around 5-10 installs per day. That's pretty awesome. So like I said, if you have an app that works for this audience, it's a great way to get some early traction.

It was a similar story on the non-Shopify side of my app. It slowly dawned on me that I had created quite a technical product (open graph image generation tool), but was aiming it at a mostly non-technical audience (Wordpress / Shopify site owners). Basically a recipe for endless support requests on how to integrate the code, fixing edge cases, etc. Not sustainable at the $20 price point I was charging.

The only people actually converting were the small segment of highly technical shopify / wordpress users who were comfortable self-serving and integrating the code themselves. But I think this is a niche within a niche.

So something had to change.

I think this post by Courtland sums up my thoughts at this fork in the road:


My app was trying to do too many things. It was an open graph image generator, for Wordpress… and Shopify… but also generated IG story size images if you wanted… and it was me designing all the underlying templates…. I could go on.

In addition to this I was getting a lot of random requests from people about generating different sizes of image, or auto-posting images to this service or that service…

Frankly, it was a tangle of different functionality and value propositions.

So I decided to do two things:

  1. Refocus on a target segment whose problems I can properly empathise with and who I’ll be much happier fulfilling customer support requests for.
  2. Simplify the core product into doing one thing well. Do less so that I can do it better.

Which brings me to the Bannerbear API.

On target user:

I’m a developer. While I can’t really empathise strongly with the problems of Shopify store owners, I can absolutely empathise with the problems of devs / dev teams. They say you need to pick very carefully and “love” your target segment because that’s who you’ll be talking to day in day out - well for me, I’m honestly super interested when talking to devs about how they will use my API!

On simplifying the core product:

Bannerbear is now an image generation service. It combines a powerful Sketch-like template editor with a simple REST API.

My job is to make the template editor better and better, and keep the API running. That’s it! You design your own templates, you send whatever data you want into the API.

The API is agnostic as to how it gets used - that’s all up to you. Create ecommerce retargeting ads, build the world’s biggest meme site, generate personalised email graphics, generate open graph images… it’s an API so you can do anything you like with it :)

Now, strategically was all of this the right thing to do?

Only time will tell. But one thing is for sure, I now actually enjoy opening my customer support inbox!

  1. 4

    Thank you so much for sharing the cautionary tale about the Shopify audience!

    1. 2

      No probs, but like I say... for the right type of product it probably does represent a really good opportunity.

      I think if your business model is more "arbitrage" based, then it could work well. Like for example, if you are able to build low-risk, super quick MVPs to make low-cost clones of existing popular apps on the Shopify store that maybe have a different target market (e.g. focused on a particular ecommerce vertical).

      It's hard to get passionate about such a model though, as an indie dev. But if you have the stomach for it, there's still a big opportunity in Shopify IMO.

  2. 2

    This is actually a very interesting post. Thanks for sharing the shopify information too.

  3. 2

    Thanks for sharing your thought process Jon, super interesting!

  4. 2

    Hey Jon,

    Exciting times! It was cool to see Bannerbear at the top of Product Hunt the other day.

    It totally makes sense that you would pivot to serve a target user that suits you better. I'm wondering though, were you also hearing from people that they wanted an API-based image generation service?

    1. 2

      During the 2 months since my last "launch" I had a lot of requests / suggestions from users that basically could only be delivered by an API. It's the simplest way to enable a ton of different use-cases.

      Nobody specifically said "I need a REST API that generates images" but rarely do we get handed these things on a plate. You just have to listen to user goals, problems, and figure out how your product can help.

  5. 2

    Great write up. It's been really interesting watching you make these pivots while slowly getting closer to finding the right market/fit for the product. Sounds like you've really simplified the product by narrowing your target customer definition while at the same time, making it much more powerful by going the API route! Looking forward to seeing how it goes.

  6. 2

    This was a really good read, Jon - very insightful. Thanks for sharing!

    I believe you made a very good move turning it into an API-focused product. Less hassle and endless possibilities. I bet you're gonna scored loads of SME and enterprise customers soon. Wishing you all the best!

  7. 1

    I love the data-as-a-service business model. Following the Clearbit model of lightweight UI and heavyweight API makes a ton of sense:

    • Cheaper to create
    • Faster to iterate
    • Better retention (API customers always stay longer than UI customers)
    • More M&A options

    Auren Hoffman wrote about this: https://summation.net/2019/06/19/everything-i-know-about-data-companies-in-30-minutes-and-one-viral-tweet-storm/

    Anyway, good stuff!

  8. 1

    Thanks for sharing. As someone who is working on a Shopify app, I agree that it can be quite painful and might not be worth the exposure it gets by being in the app store. API's are constantly changing, and the whole thing starts to feel clunky when you embed the app into the Shopify backend.

  9. 1

    Great read Jon. I'm curious, as you were seeing about 10 installs a day organically from the Shopify store, did you ever try to increase the price until you were only seeing one or two installs? and did you experiment with pushing more on the cold-outreach approach?

  10. 1

    Is this the coolest 'product' I have seen after Canva in this space? :D

  11. 1

    This comment was deleted a year ago.

  12. 0

    This comment was deleted 2 years ago.

    1. 1

      Either you’re replying to the wrong thread or you’re misunderstanding what my product does...

      I don’t host client sites in any way.

      And my API is already provided over an SSL endpoint.

      1. 1

        My bad, I did reply to the wrong thread haha. Had alot of tabs open

Trending on Indie Hackers
How do you decide what idea to work on? 38 comments Design & UI/UX takes me so long- am I doing everything wrong? 31 comments Looking for feedback on a note-taking tool focused on your personal interests. 7 comments My new self destructing notes app is on product hunt today. Would love some support. 6 comments Here's how I'm going to try to scale my tech newsletter from 0 subscribers to 10 000 subscribers 5 comments Looking for an Product & SEO/Content Marketing Partner 1 comment