What are your top ways to identify an idea as a winner? Search engines? Reddit? Spouse? Facebook? Adwords?
Look to existing products/markets.
Build a product in a category that is already well established and validated. Of course you then lose first mover advantage, but the vast majority of businesses are not entering a truly blue ocean market anyway.
Since your target audience and idea will be well established (and validated) the goal then becomes building a product that differentiates itself from your competition. This can be done in numerous ways such as via features, visual design, marketing etc.
Also, as others have mentioned, go out of your way to talk to end users. Every time we've talked to our users and customers we've received actionable feedback which improved our product.
People have used black hat SEO and the like - the kind that'll get your site banned from Google - to drive traffic to an 'MVP site' with a buy button that has no product behind it. Enough clicks and the idea is a winner. Send the people that clicked an email with a fake story explaining why it isn't ready, and get them into the funnel.
I was listening to one of these: https://www.youtube.com/watch?v=OpiVyKIJ3AM
(Can't find the exact one but they were selling for 20 or 30 million, it was two ex-finance Australian guys and the company was actually a gaming (as in videogames) clothing company) Where the founders had done this kind of thing over the course of I think 2 years until they found their product.
The guys at http://tropicalmba.com also did this for at least one of their businesses. They actually used a photoshopped an image from their competitor's site and used that for their psuedo-offering.
From what I understand the idea was made famous by Tim Ferris in The 4 Hour Workweek, but it's still pretty little known.
To clarify, the buy button shouldn't actually take CC info.
My approach (which I blogged about here) is to come up with as many ideas as possible, and then evaluate them against a rubric of 7 things:
#1 is where I try to calculate the market size, but I'm usually not too rigorous about that to be honest. Based on the interviews I've done on the site, most other people aren't very rigorous, either. Here's my guess as to why:
If you're trying to be an indie hacker and build something profitable for yourself, you only really need to make $5-10k/mo, which is almost always doable. You don't need a billion dollar market. So the size of your market becomes far less important than the quality of your product. Specifically, what matters most is that what you're building solves an actual pain point and that people will love it.
What helps the most with determining that is understanding your target customer on a personal level. Usually, people shortcut this by building a product that helps people like themselves, because then they're guaranteed to understand the customer.
Hey I read your blog post. I like the AI idea best too. I posted about something similar in the ideas thread. I'm curious why you think we're unable to handle something as simple as calendar management.
For example: "Update my schedule for a coffee with such and such at such and such place tomorrow at around 2"
It's programmed to assume 2pm for a coffee (and most things other than some kind of party, in which case it'll ask). It'll reference other calendar events for that day. If there's something going on within 3 hours, it'll use a Google maps API to check how far both locations are to eachother and make a suggestion (or not) accordingly.
As far as the involvement of human input, which is the basis of a lot of deep learning from what I understand, there's even more opportunity there. My mom audits health charts at UCSF. She's worked in the medical field since she was 20 and says that this job is the most boring she's had be a large margin because how easy it is. There is a clear market calling for automation here with a DL algo learning from their inputs. Also with the retail isle counts. You can start at a mom and pop grocery store, having people in a developing country count the items from the same images and teach the algo that way. Storewide floor count can be done in a night (as opposed to over the course of 6 months at a larger store)with an employee taking a video of the isles.
I think AI can go a long way with your calendar+email, but it's still not nearly as good as having a real life personal assistant. It mostly comes down to the UI. A real assistant can get stuff done without constantly requiring you to open an app and tell it every little thing to do. He can also talk to clients and work with their schedules. Clara can do this, too, but I believe it's actually powered by real people :D
I'm not sure what constitutes as AI, but I think something like that can be achieved without machine learning. The app would already know general times (which can change if you give a conversational command to override it - ".... and I can be available from seven to ten for this" ) you're open to meetings and it would know your prior engagements.
I think what it comes down to is just writing a lot of code and good communication with users.
I tried doing customer development interviews using the script from the book Running Lean. it saved me from going down the wrong path. Its much faster to talk/listen to a few people than it is to build an MVP.
This admittedly takes longer, but I emailed hundreds of people who were in the market I wanted to serve. Not many replied, but I had good conversations with the ones who did.
Many founders try to avoid talking to people, or more than a handful if they do. The problem is, assume a magic genie gave you the validation you needed to build your business, now what? In order to sell, you'll have to get in touch with thousands of prospective customers anyway. If you're unable to get in touch with the market on day 1, you're only going to have this problem again later on.
If you can't adequately get in contact with your target market then chances are your product will never find its market fit, dooming your chances at building something really successful anyway.
I build things which solve problems for groups of which I am a member. I have little patience with friction or lack of polish, so if I regularly use something I made, I have confidence my target audience will.
That worked well over many iterations as a designer for a corporate product in which I shipped many features that were not requested or anticipated by customers, but were quickly adopted by most if not all customers.
I expect it will also work well when I begin to ship indie products.
I really like this approach, thanks for the info!
Read all 3 parts of this, great advice , https://medium.com/@cliffordoravec/the-epic-guide-to-bootstrapping-a-saas-startup-from-scratch-by-yourself-part-1-4d834e1df8c1#.f4pmuwnyb
I've tried what seemed to be the consensus of throwing up a landing page via Instapage and trying to drive AdWords traffic to it.
Google turned around and banned my ads as they were against policy. I banged my head against them for a couple of days trying to get an answer as to why they were against policy and then gave up.
I then built a WordPress site with more content on it (http://www.notatworktoday.com/) but they still wouldn't run it.
From reading I did, the landing page > AdWords > join waiting list is not allowed by Google AdWords as you're not an actual business.
Obviously I could remove the waiting list bit and do a smoketest on taking card details but that's a step too far imho.
Any ideas?
I've heard there are Google probationary periods for sites under 6 months old, but these are just applied to rankings.
I'm not sure what's going on. I'd suggest emailing an SEO guy and asking. There must be some little standard you're missing the mark on.