Almost ten years ago, two friends of mine were developing an amazing recommendation system for movie buffs all around the world. One that could put the recommendation engine of IMDB to shame. You would have never heard of it. They shut it down months after launching it. There wasn’t enough traction.
It doesn’t work that way. Ever. For the most cases, your users don’t care what you do or how good your product is. They are all selfish — just like you and I — and there is no reason why they should be concerned about how good your product is. The only thing that should and does matter to them is whether or not your product does something that is of any value to them. And if it does do that, then the next thing they are concerned with is how frequently do they need this value getting added to their lives, and if there are any costs — intrinsic or otherwise — associated with receiving this value. The cost could be time, effort, money — anything.
‘Build it and they will come’ almost never happens, and neither does social sharing.
You want your users to be your brand ambassadors and influencers? Answer these two questions first:
Not ‘would’ but ‘should’.
And whatever your responses to those two questions are, ask yourself a common follow-up question — SO??
If your reasoning is still able to hold its ground after the third ‘SO?’, you are on the right track. If not, you’ll know it’s time to get back to the drawing board.
Pro tip: Don’t wait till the day of the product launch to ask these questions. These questions need to be ingrained from the get go. These questions and the answers to them have to be a part of your overall product strategy and roadmap.
The system was good. It could suggest you movies based on your preference, genres, map your preference with other users on the platform and accordingly suggest you the right movies. It did give pretty accurate suggestions — back when I tested it when it had just launched.
There were problems with the UX though.
They later on fixed some of those issues, but the larger issues remained. Just look at this screengrab from one of the later releases.
We fought and fought over this. I was imploring them to launch a rough cut of the product on priority, and they wanted to make the perfect product.
That chase to perfection took them 2.5 years. And while they did manage to make a product which had quite a robust backend architecture, they dropped the ball on a lot of things that matter to the end user.
More importantly, the longer you work on improving the product, the higher you start perceiving your chances of success as.
You end up making a product that is so good that you can think of no possible way it could not work, let alone fail. The expectations get too darn high. So, when the product doesn’t even come close to receiving the kind of welcome wagon you were quite certain of, it gets quite real quite fast. The euphoria turns into disappointment, and you choose to hang up the spurs. That’s what happened with them — more or less*.
Product lifecycles are long. Development takes time. Building a tribe, a community of loyal users takes an even longer time. Don’t wait for the product to get ready to start laying the foundations of building that tribe. Start laying the foundation as soon as possible.
You are a movie recommendation system? Simply set up a slack account. People can add that contact to their teams and ask for a movie recommendation at any time. Heck, go for a whatsapp number of a facebook messenger page. I don’t care if the account is not linked to a bot but an actual person in your team. At least you have got the system up and running. At least you will get to know what your users are going to expect out of the system that you are trying to build.
And as you keep on collecting that data and that user behavior, start building the product. Keep on making more and more of the process automated in increments, phasing out the human agent.
Call it agile development, or whatever you may choose to. My point is simple. The process works. We use the term agile development, but actually use it neither effectively nor efficiently. Starting from a very basic system (even if manual) and then gradually building up on top of it helps you in many ways.
You know that the system is not ready yet. Far from it. So you are aware of the fact that there will be bumps in the road. That users won’t be having the most ideal experience out of it. But you know why you are doing it, and you are simultaneously working on getting the right system in place. So, its okay.
You are not done yet.
Then comes the part of user feedback — whether solicited or derived from their behavior and usage of the current system in place.
Pay attention to how the users are engaging with your system, what kind of questions do they ask, what gratifies them the most, what irks them off. All of this is product feedback, without even having a product. Use this to fuel your product development process.
Now. When you do this, you gradually arrive at the stage where you wanted to begin with. The real product.
Your users would have been a part of the overall evolution. If you have been actually adding value to their lives, they would love you for it. All in all, everything will be just peachy!
It shut down, and it didn’t shut down.
As I mentioned the backend was quite solid. Last I knew, he was selling licenses of his algo to companies looking to incorporate a recommendation engine in their websites. Had sold a couple to companies in Canada, Australia etc.