aShop
Effortless Stores for Amazon Sellers
Only now managed to recover from the news I'm about to announce here.
It has been months since I completely shut down aShop and shelved the codebase and all my knowledge related to it.
I was unable to rewrite integration with Amazon, after AMZ announced API depredations with such a short timeframe.
But that's not the only reason to shut this project down. I've been silently dealing with bigger issues in the background all this time:
- Bad goto market strategy
- Wasting my efforts on features/things that don't directly benefit my customers
- My "founder incompatibility" with target audience (I never sold anything on Amazon)
I failed, but learned a lot during this trip. I have no regrets and would probably do this again, just better this time.
I still believe in aShop as a viable business idea, but I'm not capable of making it happen alone.
We're delivering small improvements on a daily basis. But for this update, I'll be highlighting our biggest changes within last month.
-
It's now possible to edit product description and title to avoid duplicated content in your stores. It could help you tailor your store to better match branded store experience and unique content will improve your ranking in google.
-
We slightly reorganized our database structure to boost performance of stores to make them instantaneous. It means that every page will render within 3 seconds to brings you better conversions and improve google ranking
-
Store operations will never be interrupted by unresponsive Amazon API's, but product updates could be affected by Inactive API keys (due to low sales activity) or by Amazon API downtimes. It has become less of an issue, since we're introducing different fallback methods.
As always, happy to hear any feedback.
Good day from a crew that runs aShop.co "Effortless microsites" project.
We appreciate everyone's feedback and questions. It was enlightening! We've been cooking a lot of answers to those questions -- so we'd love to present our status update.
Our biggest announcement is add-ons system for microsites. We published details about 4 premium features that you can get additionally for each site (for $9/month each). Those add-ons include:
- Your Amazon Affiliate Keys/API access
- Custom domains with free SSL
- Facebook and Twitter pixel
- Whitelabel site
One add-on is available already . We also did a lot of small improvements across the platform and eager to hear your thoughts!
Thank you
I'm building a microsite builder platform for Amazon Sellers and we're currently in beta. Have some paying customers that already use our platform.
The platform itself works smoothly already, but I've run into issues with Custom Domains.
- I have to support wildcard domains like *.ashop.co for free users (or those that didn't provide custom domain)
- Paid users can provide their own Custom Domains -- we currently support any SSL certificates that user can upload. But we plan to support only Let's Encrypt in a completely automated way (most of our users are not tech-savvy, so less hassle - the better).
And we ran into limitation at Heroku, here is an article:
https://help.heroku.com/8P5TVA4T/how-can-i-configure-multiple-ssl-certificates-for-a-single-app
- I'm using wildcard SSL, so I can't really use Automatic Certificate Management.
- ACM is limited to 100 domains, so i probably can't host all my customers on a single heroku app.
Currently I'm generating a separate app on Heroku and add SSL endpoint. I totally missed, that I have to pay 20$/m for every SSL endpoint... but my package is 19$/m.
It appears that I'm losing money on every client right now :)
So i'm looking for a better way to solve this custom domain/SSL issue. Maybe someone had experience with similar problem?
AJ from carrd.co seems to had similar issue. And he doesn't support both domains at the same time - If someone provided custom domain, then *.carrd.co stops working.
I think I can achieve the same result if I'll separate my app.
- Main app will handle all the admin + *.ashop.co subdomains with wildcard SSL
- Secondary app will be deployed to separate heroku app and will handle only CUSTOM DOMAINS. I can also create a follower read only database for it to read from, because microsites communicate through API with a main app.
Did anyone do anything similar? Am I missing any gotchas?
-
2
Funny you posted this, and thanks for making it so detailed! I actually just had to solve this exact problem for www.applight.io and if you're interested in the technical side of things, I can let you know what I did :)
The TLDR; version is:
- Create a wildcard certificate for
{subdomain}.your-app.comurl's and upload that to heroku to cover all of the base domain - Setup a way to store "Pools" of domains, like you said, up to 100 per pool (I use Rails so store it in a
CertificatePoolthat has manyCertificates) - When a new domain is added, see which pool has available space and add it there.
- When a pool is updated, send a call out to LetsEncrypt via the ACME protocol with a new list of domains in the pool and go through the process of verifying each of them (I can go into more detail on this if you want). In the end, that should return a new shared certificate that now covers up to 100 domains.
- If you haven't yet, use the "Heroku platform api" to spin up a new app, provision the SSL Addon, and upload your shiny new certificate to your new app.
- Add the custom domain to your original base app.
There's some other fun stuff to do with setting up a DNS record (I use Cloudflare) to point something like
dns1.your-app.comto your new heroku app (not the base app, but the one hosting the certificate).Users can then just add their domain, update the cname to
dns.your-app.com, and Heroku will serve the app up via your base app but pull in the certificate via the second app. You can spin up another app each time you get 100 domains.At that point, you're making $1,900/mo and spending an additional $20 for every 100 domains :) I like that profit margin.
Let me know if you need/want/care about more details! But even the TLDR was long. Learned lots about SSL, more than I ever desired to :) There are some gotcha's hidden in there too, I considered writing a blog post (and even spinning up a paid app to do this, since it was so annoying haha) and may do it at some point but always here if you need help implementing it!
Mike
-
1
great! thanks for your reply.
You just confirmed that my idea should work!
Would love to hear more about ACME protocal. It sounds like you've built a custom solution and avoided ACM that heroku provides?
- Create a wildcard certificate for
We've been working through all the feedback that we received during our private beta and worked closely with couple of new customers on our platform.
Since most of the people are not quite familiar with microsites and how those could be used, we often heard requests for a free/trial package.
To be frank, we completely discarded free package almost immediately after we started working on this project. But we had to sit down and reconsider...
We've decided to launch a free package, that will include most of the benefits that we already outlined -- but without cart support and without custom domains.
Work on this is already ongoing and we hope to release this update soon!
Unfortunately, we failed to frame our project properly and explain value we bring to customers. We will not refer to our project as a 'store builder' anymore and will use a 'microsite builder' term. 'Microsite' brings a clear message.
We're also reducing our pricing packages to micro levels - now $19/month reduced from $39.
I'm happy to announce, that my main project is launched as public beta.
https://betalist.com/startups/ashop
I'm a single founder, but this would never been finished without all the support and help I received. I'm very grateful to have such a great family, friends and first customers
After a lot of testing with two stores, we've finally decided to open the flood doors...just kidding...
We submitted Ashop to betalist.com and got accepted. We would try to pick next 10 stores for our platform and work closely with them.
Same customers, polished platform, simplistic, performant and effective.
Biggest (and most requested) feature: custom domains with ssl support.
We managed to go through black friday and cyber monday without breaking a sweat.
We built horrible looking admin panel, so our clients could manage important aspects of their store. Probably because of that, we ended up with second store on our platform.
We did another soft-launch -- to see how the platform will perform in real conditions. The biggest revelation was to understand that crucial features are still missing.
8 Comments
It was a quite long story for more then 3 years as I can see.
There is another indiehacker here who does (seems) the same busyness (though in French only) and he seems doing quite well. So, you may rethink, revise and remake it after awhile. The experience is what has real price in your case.
Thanks!
I did write down all the lessons and plan in case I want to start this project again.
But it was worth it for experience alone, as you said!
Hats off to you and your efforts. Lessons learnt are extremely valuable for the next startup idea. Don't worry, it will only get better as this is a marathon and you are just at the beginning.
...keep grinding!
... will be filling an indie hacker page with a new project I'm working on!
Thanks for encouragement! Means a lot
Shutting down takes a lot of courage, even more so to announce it in the public. I think most of us just let our projects sit on a shelf indefinitely, thinking "maybe someday..."
Best of luck with your next thing!
Thanks. These words mean a lot!
Sorry to hear that, man. I have a backyard of failed startups too. Just to name a few: Todayter, Voolkan, New Century Marble.