How we built disruptive innovation from ground up
What enormous joy I feel today introducing our new application — personal real estate agent Maakler in your pocket! It is a real estate application, currently for the Estonian market, which automatically calculates the property value for every apartment building in Estonia, brings together most of the real estate ads in one convenient Google Maps-like application and creates a pre-market for people who are interested in properties even before the public ads are put up.
In the spring of 2015, I met Andree Raid, the founder, and CEO of a real estate agency Raid & Ko. He was curious to hear what I was doing in Rubik’s Solutions and we started discussions over the opportunities for innovation in the real estate market.
Andree has been in real estate industry from the beginning of the 90s and has a great understanding of the whole sector. He had also been building a real estate portal himself but the portal wasn’t that different from its competitors. So we started thinking if there is a way to create new, significant value on the market by redefining the typical real estate advertising market. By analyzing some of the key obstacles in the market I also learned that the evaluation of properties is a huge hustle for the end-users and people pay several hundred euros just to get an understanding of the price of their home.
For me that sounded ridiculous — in today’s information society, people are still forced to pay hundreds of euros for a piece of information. It is like “before our time” people went to the car seller and the car seller always had more information than the customer. That enabled the seller to manipulate the price and if the customer was “dumb enough” one could earn a ridiculous sales commission. Today a similar thing is happening in the real estate evaluation market. The most extreme examples include some of the biggest banks that actually do the estate evaluation by themselves. They will say how much your apartment is worth so they can offer you “the best loan” — isn’t that crazy? This is a clear conflict of interest and should be illegal, yet it isn’t.
Where to start?
The key element for building disruptive innovation is to forget everything about the solutions currently available on the market. In the case of real estate, the Estonian web portals have had no significant innovation from the beginning of the nineties. When Internet came, the typical yellow pages were just taken to the web, but in essence, they had the same logic that they had before the web — a huge pile of real estate listings to go through, accompanied by annoying advertisements. The search engine was really the only and the most obvious innovation added during the transition from paper to web.
I started off by sketching some plain simple mobile app prototypes — wouldn’t it be great if you could click on any house in Estonia, just like you can do in Google Maps, and opens a window of information, stating clearly how much this property is worth. Simple!
During the time I happened to meet an IT-guy who happened to work with a similar tool to calculate the average rental value of different properties. Back then, he was serving his last months of the military service and was working with the application to keep his IT-skills sharp while in the military. We soon decided to join forces to make a joint proposal for the client. But not only — this guy is Kaspar Triebstok, today the Technical Manager of Rubik’s Solutions and my business partner.
Meeting, negotiations, and idea formulation took over three months and in June 2015, we agreed to go forward with the project. Our initial estimation for the development was four months and we brought in another dedicated developer. It was a busy summer.
The key feature of the Maakler application is the automatic property evaluation — our idea was to put a price on every property in Estonia, to fully automate property evaluation. In order to generate two numbers for every building in the country — property evaluation per square meter and rental evaluation per square meter, is, however, a huge hustle.
For the property evaluation, we first needed to get connected with The Estonian Construction Registry through the X-Road which is the backbone of the Estonian information society. Taking statistical information from The Estonian Land Board and combining it with the information from The Estonian Construction Registry, we got the base information for our algorithm. From there on, the magic began — combined with some machine learning and the ability to process huge amounts of data, our algorithms started processing the data, increasing or decreasing the estimations based on the same factors that the professional real estate evaluators use (designing the algorithm also meant working countless hours with a professional real estate evaluator).
“Oh, of course, you have an app for it!” said an Englishman when he asked about the property prices in the area and I opened Maakler
And this is just the property evaluation part, rental evaluation required a whole different algorithm which is not based on deals done in the past but is based on the public rental prices currently on the market. If you are buying a property, you want to know the actual value of the property, not the price that it is being sold for. But if you rent a property, the most accurate rental price is the one that other similar properties are currently offered for, in that region.
Problems, problems and more problems
Summer went by quickly and although a lot of progress was made, by the end of the four months, our key assumption about the automatic evaluation was still uncertain. Connecting to the X-Road proved to be a much more complicated task than anticipated — after countless obstacles, hundreds of calls to different X-Road representatives and months of work, we finally managed to get the connection, thus proof of concept in the end of July… July, the next year, 2016!
Additionally, in the fall of 2015, the previously mentioned real estate portal which was being built by a small development company, and was meant to become a supportive web interface for the application, encountered its own obstacles. The company building the site announced it will be closing its doors and despite its kind offer to finish the ongoing projects, it soon became apparent that the site will never be ready. That meant we had to build additional features in the application, features that would otherwise have been covered by the web application.
Months went by, initial resources were drained and obstacles were mounting — our partner and investor announced that he is not interested / able to invest further in this project — as he had also invested and lost resources with building the web portal, it was a reasonable decision. Also the agreed four months with the external developer ended and he left.
As the technical plausibility of this project was still not proved, it was mostly Kaspar’s decision if he is willing to put additional working hours on the project. It was his determination, curiosity and desire to not give up that kept this project going. Nearly every month we stumbled upon a problem that made us consider scrapping the project, but every time it happened we also managed to find this one thing, that we could still try out, which then happened to lead us to the next level.
Although I didn’t approve any front-end application development before we hadn’t proved our key technical assumption of the plausibility of automatic evaluation, Kaspar still continued to build the application during waiting behind several X-Road issues.
The beginning of the magic
In the end of July 2016 when I was enjoying my peaceful vacation, Kaspar finally messaged me: “We have the connection!”. Indeed, we finally managed to get connected to the X-Road, and for the first time actually try out our algorithms. Like magic, it worked! Most of the apartment buildings had an estimation on them, and not only, after some research, the numbers also proved to be quite accurate, what a joy!
This is when the work really started. We allocated all of our efforts as well as resources on improving the application, preparing the marketing & sales, getting ready for the launch and testing, testing, testing in every corner of Estonia.
The search engine was really the only and the most obvious innovation added during the transition from paper to web
By this time we had put in majority of the resources for building this application and the initial shares of the company (to be founded) had to be reorganised — we became the major shareholder (over 70%), raised some additional resources from the current investor and brought in one additional investor Manimal Ltd to fund the marketing campaign and help us with design.
Past two months have been crazy busy to get everything up and running. Regarding the application, there is a common saying in IT, that the last 10% of the application takes 90% of the time and indeed the bugs and small glitches in the application seem to be endless (as always).
Nevertheless, I believe Kaspar has done a great work with this application, he has significantly overachieved what should be possible for a guy who was basically a junior developer in the beginning of this journey and for whom it was his first ever Android application.
Let’s get crazy
“Oh, of course, you have an app for it!” said an Englishman lately when he asked about the property prices in the area and I opened Maakler to check it out. It does seem kind of an obvious thing to do now and I have truly enjoyed using this application myself while moving around the city and traveling the country. It’s been a game-like experience to check the prices of the properties nearby. It really feels nice to own some kind of information that others don’t have. But the real, enormous joy, for me, comes from giving it away today, to everyone, for free! Please, enjoy it!
by Erik Ehasoo, Business Developer, ICT Futurist & Opportunist
PS. Having had extremely limited resources on building this application, we still have a lot of work with improving it. Please help us change the real estate market together and send us all kind of feedback and glitches that you find in the application. We really appreciate any help!