I’m thrilled to tell you that we’ve successfully closed a Series A funding of $1.5M. The round was led by Helion Venture Partners, one of India’s largest funds and investors in MakeMyTrip and Redbus. The sweetest part of the deal is that Ashish Gupta, an old friend and mentor, has joined our board. Existing investor Blumberg Capital also participated in the round.
Now travel is a tough segment. And travel planning is even worse. It is a fundamentally hard technical problem and often the rewards accrue only to those who manage to be close to the transaction. As a team, we’ve lived, breathed, ate and dreamt travel planning for over three years now. We threw out our first iteration and now after months of staying under wraps, we now have a product that I think is a technical and design tour-de-force.
With the typical humility of an entrepreneur, let me tell you 3 reasons why Mygola is unlike anything you’ve seen before
Reason #1. We let travelers do in 15 minutes what typically takes 6 weeks
Travelers take weeks to go from the inception of an idea to making final bookings for a trip. During this period, their emotional states change wildly. At the start, they want to be inspired and excited about the trip. Pretty soon, they become more utilitarian – trying to figure out a rough outline of what their trip might look like. And when booking, they’re hyper-focused on value and trying to stay organized with their logistics.
Typically, this plays out over 4-8 weeks. Something like this.
Unfortunately (for a startup), that means that travel planning is a long, leaky funnel and you’ll lose users all along that path. That, alas, is the classic definition of startup death. We debated this for a long time and concluded that our best bet was in changing consumer behavior (always a fool’s errand, as any startup guru will tell you!).
That we would try to compress 6 weeks of planning into 15 minutes.
Now put on your design hat and think what that means. You have to inspire users about a destination, so perhaps a very visual metaphor would work. You have to help them make an informed choice about individual places they’ll see, so the interface has to be rich with data like reviews, ratings, tips etc. You have to then give an intuitive interface to fine-tune the plan – add a place here, an extra day there – that gives them a sense of control and flexibility. And finally, you have to give them a single page to book everything that seems credible and trustworthy.
That’s the emotional transition we wanted users to go through. And that’s why you’ll see interfaces such as the ones below, each of which has been simplified, updated and re-imagined for months now to make the whole experience flow.
Reason #2. We’ve solved some really hard tech problems
When we interviewed hundreds of users who travel independently, we noticed a few unusual behaviors.
Users would hunt for a good “default” trip to start with. We found them copying trips from articles in the NYTimes, or from blogs and even from tour operator websites. So we thought we’ll aggregate all such itineraries under one roof. Turned out that was easier said than done. We had to write custom text-mining software that could parse these articles, extract structure and guess the path taken by the traveler. We had to teach our algorithms to understand esoteric things like these:
- A traveler might say that she had “sausages for breakfast in my camper van” and ended up “in Wellington for a hearty meal“. Our algorithms would interpret that to mean that they reached Wellington for lunch, probably by driving, and then backtrack to figure out what place they might have been 5-6 hours ago at breakfast
- A traveler in New Zealand might mention in passing that the owner of the lunch place was from “Munich“. Our algorithms had to figure out the probability that the traveler might have taken a quick flight to Germany for lunch, and learn to ignore such misdirections
In the picture below, you can see how our algorithms would convert a New York Times article (left) into a structured representation (right).
There’s something else that’s quite powerful about having such a vast collection of structured itineraries. We’ve been able to mine them to come up with “circuits”. That is, we can recommend what your 5-day trip to Austria should look like based on what other travelers have done in the hundreds of itineraries we’ve found for Austria on the web, even if none of them was 5-days long.
You can bias these circuits with interests (eg, I like architecture) or with certain cities that are a must-see. When users modify these itineraries, it makes our database even richer. This is the kind of Big Data analysis that can make this platform truly powerful and predictive as time goes by.
Users would create a spreadsheet, and then use Google maps to plan their days. They would put their wishlist in a spreadsheet, note when things open/closed, how far they were from where they were staying and then figure out in what order to visit them. In computer science, this is called the Traveling Salesman Problem (TSP) and is a classic example of an NP-hard problem. Suffice it to say that when a problem is “hard” for computers, good luck solving it manually!
You can see one such attempt below – this is an actual user’s plan for Germany.
We solved a variation of this problem through a unique application of TSP. The algorithm takes into account open/close hours, how far things are from each other and even when they’re typically visited.
And users were definitely not having fun while doing all this. We’d find them with twenty browser tabs open, cut-pasting things from one window to another, shuttling between Google, Tripadvisor, forums and back to Google. Some of these notes would move into a spreadsheet open on their desktop.
It was evident that they seemed to be having as much fun as filing their taxes!
On Mygola, we wanted the experience to transport you to your destination even as you planned your trip. That required us to build large crawling, indexing and ranking infrastructure that would fetch evocative images, relevant videos, stunning panoramic views and authentic tips left by travelers from all over the web.
Talking of inspiration, where would you rather go? Here?
Reason #3. We have an unusual approach to an unusual problem
We take pride in our unique ability to combine deep tech with massive-scale curation to solve the hardest problems in travel planning. Every day, we have hundreds of travel enthusiasts from around the world use our Mechanical Turk-style platform to curate images, videos and a host of other metadata for every travel item in the universe. Let me illustrate with an example.
In our effort to let users viscerally experience their destinations even before they get there, our design gives fullscreen photos a special place of pride. As you can imagine, it’s not an easy problem to find a fantastic image for half-a-million items! This includes museums, cafes, street fairs, amusement parks, nightclubs and what not.
Turns out, it is the perfect example of how we used our “Robocop” approach.
Imagine Taj Mahal. Our algorithms go out and search creative-commons image databases like Flickr and Panoramio for pictures of the monument. They have to distinguish the one in Agra with the one in Las Vegas.
We assign a probabilistic score and then do image analysis on the shortlisted ones to see color saturation and composition, all automatically. Even after this, we might be still left with 4-5 images. Unfortunately, this is not good enough – afterall, we need just one. These images are then shown to one of the hundreds of individuals who log into our platform every day.
They would see something like this.
We have sophisticated algorithms to figure out if someone’s gaming the platform. And at the end of this arduous process, a single image emerges that is creative-commons licensed, high quality and attractive.
This effort gets repeated for every other metadata – videos, open/close hours, tips from travelers, panoramic views etc. This is what has allowed us to create one of the largest databases of structured data about every place on the planet.
And if you’re still hungry for more, check out how we solve the problem of local transportation in a similar way.
We’re just getting started
I hope I’ve convinced you that we’re a new kind of company.
Our DNA is technology but we’re happy to throw thousands of people at a tough problem too. We’d love to show off all the cool tech behind the scenes but are also acutely sensitive to how important design is to keep the product simple. Our team is mostly in India but our founders and investors spend enormous time in the US. Inside the company, we’ve combined hackers with Ivy-Leaguers, journalists with backpackers.
Travel planning is an unusually complex problem. And we believe you need an unusual company to truly solve it.
Latest posts by Anshuman Bapna (see all)
- 3 Reasons Why Mygola is the World’s Best Travel Planner - October 17, 2013
- A Curious Means of Conveyance at a Fabulous Price - September 26, 2013
- Native App vs Mobile Website – a Founder’s Perspective - August 14, 2013