How NoBroker is Building the Ideal Broker with an AI Brain
It’s no secret that real estate in India is a crummy and unwieldy industry people prefer staying away from as far as possible. It’s also a gargantuan market churning tens of billions of rupees in revenue just as broker commission. And when you set out to challenge its archaic methods with a simpler, modern, and more affordable solution, you invite mob attacks to your doorsteps.
Against all odds, though, NoBroker is flourishing today in an undisclosed office location in Bangalore, India with 4.5 million customers and 105,000 properties across five cities. Its biggest accomplishment, however, is none of that. It’s the fact that it has just in the last financial year saved brokerage fees worth a hundred million dollars. So how did NoBroker break into the onerous real-estate sphere and managed to stay there with an exponential growth? Turns out, the answer involves a lot of machine learning.
But first, a little background.
NoBroker kicked off a little over four years ago when its three founders — Akhil Gupta, Amit Kumar Agarwal, Saurabh Garg were still in college. Akhil built the website, Saurabh took care of the marketing, and Amit headed operations. Like most startups, NoBroker too stemmed from a personal incident. Fed up by misleading brokers, their unreasonable fees, and substandard properties, the trio had a light-bulb moment. And just a couple of months later in March 2014, the NoBroker portal was born.
For the first two years, NoBroker’s purpose was quite simple — connect owners with tenants and keep brokers out of it. Propelled by positive feedback and a lot of bootstrapping, NoBroker continued to grow and about fifteen months later, landed funding.
NoBroker’s objective was, though, suffering from sort of a tunnel vision. The founders had yet to perceive its true potential. Pesky brokers were just a part of a larger real-estate muddle. The chaotic documentation process, the relocation headache, furniture, you get the idea.
So when users began getting accustomed to a convenient platform like NoBroker for discovering properties, they consequently also became more disgruntled with the rest of the tedious process and felt it as taking a step back. The company did notice this behavior. Either through suggestions by customers or from their own experiences. Around the same time, it also came to terms with the fact that real-estate is a slow business.
The Google of Real-Estate
“It [real-estate] is extremely slow. It takes a lot of time, unlike food or shopping. You cannot nudge a customer. Obviously. You cannot force the customer to change your house by house. So it’s a non-discretionary decision which happens only when the customer needs.”, said Akhil Gupta, CTO, and Co-Founder, NoBroker in an interview with TechPP.
NoBroker had to become more involved to make its presence felt. And it did just that. The startup added a host of utilities and tools to its platform for lending the buyer a hand in what comes after they’ve found their next home.
It begins with the rental agreement. Instead of you going through the hassle of Indian government offices, NoBroker assigns an agent (not a broker!) who takes the agreement fee, your signatures, and gets it done. Similarly, you can hire packers and movers or pest control or just about anything else. These, however, are not NoBroker owned branches. Being a technology company, NoBroker has essentially allowed vendors to integrate their services on the NoBroker platform.
So what happens is say you’re a NoBroker user who has used the platform to find a new house. You’ve shifted to a new apartment and you’re in search for a company that does pest control. You head over to NoBroker, select that you need pest control, pay the fees, and NoBroker assigns the task to a partner.
By adopting this approach instead of simply letting any pest control firm to list their service on the platform, NoBroker is ensuring customers by taking responsibility for the entire process and putting its brand on the stake. In addition, NoBroker also has a visitor management and community app for apartments called NoBrokerHood. That contributes further to NoBroker’s ultimate goal which is to be, as Akhil puts it, the Google of real-estate and that does make sense considering how vastly they’ve expanded themselves in the last two years.
The Ideal Broker
The reason why NoBroker has been able to scale so expeditiously is automation. Since the inception, a flurry of machine learning algorithms has powered NoBroker because of which today it likes to call itself the “Ideal Broker”.
“Why we have been able to scale it is that we have used the technology beautifully. The technology and the data are the two backbones of this company. Whenever and wherever we see that there is a chance to automate or bringing the technology to solve the problem. We go ahead and do that.”, added Gupta.
One of those algorithms is designed to justify the startup’s name by blocking brokers. Generally, real tenants would browse listings in just a handful of areas at a time. A broker, on the contrary, would go through heaps of properties since it’s collecting owner contacts. The framework works by basically sifting such user patterns to prevent the latter from accessing the details by not sending it the one-time code when the Contact button is clicked.
But let’s circle back to the concept of Ideal Broker because that’s what has largely propelled NoBroker’s growth.
Hypothetically, you would want a broker who learns every one of your needs, knows the locality and suggests properties accordingly. But that doesn’t happen in real life. It’s not even close. NoBroker wants to make that a reality and it’s doing so with an artificially intelligent framework.
To do that, the algorithms evaluate a bunch of scores for each property which are based on several factors such as how far it is from your office, the included amenities, nearby hospitals, public transportation, entertainment landmarks, and more. Therefore, when you look for properties on NoBroker, the search engine takes into account these scores and sorts the results accordingly.
“They’re our own 15 attributes of the property. So higher the mass that means the property is hot. So based on that and the requirement the customer has, we create and we show the higher mass properties to the customer. So that’s how the recommendation happens.”, explained Gupta to us scrupulously.
The Ideal Broker would, of course, know the fair price for a property too and NoBroker has figured out that as well. Through the same attributes which influence the livability scores, NoBroker’s algorithms can also estimate the right amount for a listing. It’s shown both to the customer and the owner too in case he’d like to alter the cost he has proposed while posting.
There’s so much happening behind-the-scenes on NoBroker but when you log on and search, you’re not overwhelmed by all of that. The only thing you face is a seamless experience where the platform more often than not knows precisely what you’re looking for 99.9% of the times, if you ask the startup’s CTO.
Sure, there’s also the question that these algorithms can malfunction and the owner can end up losing potential clients. But Gupta ensures us that that rarely occurs and since it’s an always-learning technology, it continues to improve over time.
It is quite ingenious how NoBroker has overcome common real-estate hurdles by sewing together a network of machine learning frameworks. Technology is what has defined NoBroker because unlike a handful of its competitors, it has conjured up an ecosystem of services to stay relevant even after the user secures the listing.
“We now have the Playbook Ready”
NoBroker, all these years, was largely dedicated to building a real-estate platform and figuring its numerous complexities. In the coming months, however, NoBroker’s journey will revolve around expansion beyond five cities. “We now have the playbook ready and we should be able to cater to the top 50 cities in India”, said Gupta on being asked about the startup’s future plans. And of course, that also includes keeping their office location off Google Maps.