The decision to launch a raw product was traditionally received with skepticism by mainstream users.
At the end of 2019, 5 months after the start of work, we launched an internal system with defects in processes, “non-intuitive” interfaces, and with migration errors. All these 5 months, an exclusively backend has been designed, which can evolve. The risk was that the elegant backend was not visible to anyone and they would ask me - "And what did you do there for so long if people say that nothing works?"
At the time of launch, there was a clear understanding that there would be more than one failure. That the data is like an X-ray, will bring to the surface all the hidden imperfections of the processes. That there will be a flurry of calls and general discontent.
“… And there will be many more comments that we cannot even foresee,” I told the frightened team, without a shadow of a doubt pressing the start button.
And people said:
"Nothing works for us!"
“You haven't collected the requirements well enough!”
"We must discuss again!"
All I needed at that moment was the trust of the management. Because user dissatisfaction is temporarily like March snow and will go away along with defects, and we will continue to live with a reliable scalable system.
And so 2020 began, in the first quarter, the main defects in critical processes were eliminated, the tension subsided and the car started moving. And then - you know why - a completely new functionality became in demand. Which was implemented quickly, and provided the conditions for effective operations in the new reality. By the end of the year, the system already had several million transactions! “The system cannot be used”?
Did we know about the upcoming changes in 2019?
No, they just took a chance and made a bet on the “invisible” that we were sure of - well-thought-out internal logic, delivery process and team interaction in order to be able to roll out ANY (and not only well-known) functionality with a progressive decrease in time-to-market.
Would we be in time if we were afraid to release a crude product in November 2019?
If we waited for demands and listened at endless meetings how users agree with each other - we would still be there.
Why get together and argue about what will work, if during the same time you can do an A / B test and see? Why figure out what will be more convenient for the user, if you can start it and he himself will say what is convenient for HIM and in what design? Technologies and instruments now - immeasurable, all that limits us is fear and our own contexts. I wanted the system to become a tool (!) Of effective processes, and not just reflect what is.
Can you call this a success?
I am in no hurry to evaluate. Time will show. Even though now something is already visible. Colleagues objected that there are best practices, and UAT testing, and different conditions, and such a trick would not work in a bank. And no, I do not support thoughtlessness and immature decisions, and I do not advocate investing in everything, I am for a practical iterative approach, which has never let me down. Especially when high level of obscurityand opinions are based on hypotheses, not years of experience.
And for added credibility, I always quote my favorite quote: “If you’re not ashamed of the first version of the product, you entered the market too late” (Reid Hoffman, co-founder of Linkedin)
Too late, because they did not know what they should have known in advance. Because they didn't start earning. Because there are other smart people who gain invaluable experience while we are guessing.
This is the case when time plays against us.