The MVP is ready. I have tested it in the wild and Youmami is able to provide recommendations as I envisioned in the MVP post. However, I am finding it very difficult to start the campaign and "launch" the bot.
There are a few things contributing to this:
Asking users to sign-up via an instagram post is cumbersome. Sign-ups should be as easy as possible and asking users to (1) make a new post, and (2) include the #youmami_bot_subscribe is too much. If someone else would ask me to do that I wouldn't do it. Especially if the request comes from a spammy bot.
The MVP is raw and basic. The recommendations are ok but many false positives, and there is no automation as I did not want to spend time setting up virtual machines and environments.
Asking for donations is a really bad idea. My original intent was to monetize the bot, not to ask for donations. I should have been firm on this.
I have not launched my campaign. As a result, I am late by 1 week according to the original plan for #costarting.
To address the two concerns above, I did the following:
Created a basic sign-up form on the blog front page: https://canedoyoumami.wixsite.com/website. This will hopefully give the potential users a familiar interface. I get the subscriptions via email and I can manually add them to the database for Youmami to give the recommendations.
Without users, or just a few users, automation is not critical. I will run the recommender once a week from my personal computer. I will add the new subscriptions to the database manually. This will allow me to focus on further improving the functionality and add automation step by step.
I linked the Paypal cohort's money pool in my landing page: https://www.paypal.com/pools/c/8vK1qzC0jW. However, I will not ask for donations. I think this is a dumb idea. I'll give a free weekly subscription to the early adopters for 1 year.
With a more reasonable plan in place, I feel more confident to launch the campaign and create awareness. In retrospective, I should have started with the campaign and then implement the basic functionality. This would have allowed me to validate early and refine the idea in the available time.