[askopress]

It's because most people run WordPress on shared hosting with little to no SSH access and even if there is SSH access it's not like I can simply let my plugin install imgmin and pngquant, that's why most of the compression plugins that do not use an external API for image compression use something like ImageMagick which renders poor results. So yes, I would say the first. There is no viable method to let the plugin do everything - including the compression, if you want quality compressing, that I know off anyway.

[asteadman]

Interesting. I suspected this was the case. I've been contemplating writing a wordpress plugin, and although most of it would be bog standard templating, I would like to be able to do some simple image processing. Being unable to shell out an external binary is unfortunate. Tangentially, Can you comment on being able to depend on the GD module being present (I'm guessing no)?

Based on your new business model, I suspect that you did not have much success with the pay-as-you-go approach? Like you, I'm contemplating having an up-front fee, but having to support said users with an external API for all time would suck. I was hoping to do it all on their end, but that may be more trouble than its worth.

[askopress]

Most shared hosts probably have GD present, but as with anything, there are those that do not. As for how high that percentage is I have no idea. Think of your target market and if it's for not-regular-folks (as in more web development focused or so) then probably GD is present as more knowledgeable people choose their hosting service more wisely, or so I'd like to think, but small company website owners, fashion bloggers and what-not probably choose by price, not quality.

The pay as you go service failed because I offered 1000 compressions for free upon account registration and people took advantage of that, more often than not they would simply use their 1000 compresses and sign up a new account with a new email and nobody bought any compressions because well, seeing how quickly people abandoned the use it's likely that they only ever needed to compress less than 1000 images.

This way to even compress any images they need to purchase the plugin. It has no barriers like account creation (which as far as I can tell people don't like to do with plugins, they'd like to install and go). As for how it will work in the longer perspective I've no idea, I've done my best to optimize the API for high number of requests and will most likely release simple addons (paid) to the plugin which will enable GIF compression, compress-while-you-upload feature and whatever else I can think up, but that, too is just a one-time price so once the user has all of those I will not be making any money from them anymore.

So, if you're asking me if $7 for unlimited compressions work in the long run? Well, if I can run it for tens of users with a simple $10 Digital Ocean droplet, sure, why not. The reality however is that I most likely, depending if it will get any traction or not, have to scale up the DO droplet up which will cost me more, increasingly. There are quite a few ways to go about making money with it and I suppose I'll test the waters as I see fit, but even if it makes no money I don't mind much because I use it for all my client' sites and paying a server fee is much, MUCH cheaper than using the competitors (TinyPNG for example).