Vested MasterClass with Utsav Agarwal, Investor, ex-Expansion @ Glovo & Uber
I am really excited to be sharing this week’s edition of Vested’s MasterClass with you today. We will be speaking with Utsav Agarwal. Utsav is an entrepreneur, investor, and operator. He has extensive experience in building teams, as well as launching and scaling B2C products and services in Asia Pacific (APAC), Europe, Middle East and Africa (EMEA).
Utsav led Uber’s Indian and South Asian expansion, spearheading Uber’s most complex expansion in Dhaka, Bangladesh. Utsav was also the Head of Expansion for Glovo for EMEA and APAC, where he was the 1st employee that built out the expansion roadmap and led the expansion across 9 countries. Note that Uber is the global leader in ride sharing and food delivery, and Glovo is an on-demand courier service that purchases, picks up, and delivers products. Utsav is currently an investor in multiple startups.
Introduction of any new consumer product that requires the user to change his/her habits is typically very challenging. When you launched Uber in India in 2013, smartphone penetration was nowhere near as high as it is now. On top of that, India is a very trust deficit environment, and with ride sharing, you’re asking users to jump into a stranger’s car. I am really curious about how you and your team overcame these two challenges (introducing new habits and trust)?
The growth in the early few months, for Uber in India, was really slow. It took us a while to turn Uber into a ubiquitous brand for both driver partners and the users.
Let’s take a step back for a moment. What Uber started as in the US or India is very different from what we see today. At launch, it was a premium black car service in the US, priced at 2x of taxis. Similarly in India, it was a luxury vehicle-only service. At that point, we knew this would not scale, but would help us create the initial buzz, and solve the initial habit and trust issue you spoke about, because people want to ride in luxury vehicles. Also, at that time, Uber only supported credit cards and had no support for wallet or cash. By virtue of that, the total addressable market was minuscule. On top of that, we did not have the driver partner platform (app) on Android, so we were doling out iPhones to drivers so they could drive on the platform.
What I’m trying to get at is that when Uber launched in India, it was not thinking of scaling or achieving hypergrowth in its first year. It was trying to establish a brand and solve the chinks in the armour, and only later look at putting the foot on the pedal and growing rapidly. So going back to the challenge you mentioned, creating a new habit for Uber meant feeding a lot of the early traction through promotions & marketing to showcase the convenience of the platform. We knew from experience in other cities that once someone experienced Uber – once someone experienced the first ride – that it was very easy for the user to get hooked on the convenience of tapping a button and requesting a ride.
As for trust, it’s still an ongoing process. It is not solved completely. I would argue that ride sharing is one of the safest modes of transportation. However, a lot of people would argue against that. When we launched in India, we emphasized repeatedly on checking the driver’s picture, license plate number, and matching that against the info on the app. We also implemented a two-way rating system to help us eliminate the bad actors.
In the last couple of years since I’ve moved on, I have seen an increased emphasis on the SOS button, on thorough background checks, and on reducing the incident response time by placing support teams across the globe. These help build trust in the minds of the users and driver partners.
As a follow up on that, how do you know when it’s time to transition from slow growth to rapid acceleration?
There were a lot of indicators. Once we had the Android driver app and Paytm wallet integrated, we felt that the platform was ready for the broader market in India. That was when we decided that it was time to expand aggressively. In most cities, we were operating with only one product – UberBLACK, and as we got closer to the wallet launch, we launched cheaper products – uberX and uberGO. And since India is a very price-sensitive market, this worked wonders!
Now, this approach might not be suitable for all markets. India was the first of its kind for Uber. In 2013 / early-2014, Uber had launched in Singapore and Hong Kong, which had much more in common with the US and European markets; so Uber’s product was better suited for them. In India, Uber took its time before deciding to go mass market.
On a separate conversation, you mentioned that Bangladesh was the most difficult market you’ve worked in. How was that different than India?
Bangladesh was not only the most difficult market I’ve worked in, but also the most operationally complex for Uber to launch in. In a lot of ways, Bangladesh is similar to India, but exponentially harder in every possible way you can think of.
First, let’s look at the ‘ease of doing business’ ranking. India hovers at 70-80, while Bangladesh is ranked at 178 or so. So we knew from the start that it was not going to be easy.
On top of that, Bangladesh was a cash only market. The usage rate of electronic payments was very low. This was very new to Uber at the time. Uber only launched ‘cash’ in the latter half of 2015, and we entered Bangladesh in 2016, so the cash part of the product was not quite ironed out.
In terms of the challenges that the country threw in front of us: it had no clear provision in the motor transport law for commercial use vehicles. Most people would think, “Why would you care about that?. Wasn’t Uber the type of company that would bulldoze its way into a market?” By 2016, as Uber had gotten bigger, our approach had changed. We engaged in healthy dialogues and sought permission from the government before entering a new market.
In addition, the banking infrastructure in Bangladesh is as archaic as it can get. It was nearly impossible for us to open a bank account, and international remittance into the country had zero visibility. This was important to us, because we had to pay our driver partners on a weekly basis, and if the driver partners did not get paid in time in an emerging market (many of them live hand- to-mouth), we had no business.
There were also a couple of external factors. One that people think I’m joking about: Bangladesh Bank Cyber Heist, where $101 million was siphoned out of the account owned by the Central Bank of Bangladesh in the US Federal Reserve Bank of New York. The illegal transfer was supposed to be for US $1 billion, but suspicions were raised by a misspelled instruction. The second was the terrorist attack in Dhaka in July 2016, which made it even harder to approach policy makers and central banks, and built the narrative around this disruptive service.
Identifying and recruiting top talent is difficult. When you were leading Uber’s and Glovo’s international expansion, you had to hire key talent rapidly in these new regions. How did you tackle this? Can you share some frameworks around identifying and recruiting talent?
This is actually an interesting question, and a very undervalued aspect of expansion. I think it starts with researching startups in a particular city and reaching out to them through cold email or LinkedIn. Usually startups are very proactive in connecting and responding. Whenever we think about launching in a new market, we do a greenlight trip to study the market, and during this trip, we try to make sure we really engage with the startup ecosystem and meet as many startups as possible.
This approach serves two high-level purposes for me. First, you get to hear the challenges that startups are facing on a daily basis. What’s working and what’s not. And most importantly you get the on-ground reality. What you hear from this community is very different than what you hear from expensive consultants – as the latter do not have their ears to the ground. Second, you build a network of startup founders and early employees. So if you end up deciding to launch in that particular city or country, these contacts will become the initial node to expand your network further. Many will share the job postings on their social media channels, which will expand the reach of that JD to local startup enthusiasts – who you otherwise wouldn’t have known. I also use LinkedIn extensively to scour profiles of people from hot startups and reach out to potential candidates directly.
It is actually counterintuitive, but I strongly believe that the best talent is not actively looking for an opportunity. Instead, we need to identify them, reach out to them, and sell them on why they should join your startup. It is like reverse interviewing. They are interviewing you to determine if they should even apply and go through the recruitment process. Why should they join you, what’s so cool about you, what are you offering them? At the end of the day, hiring is a function of the brand value of your startup.
I think hiring in the initial days is not something that I will outsource or depend heavily on a recruiter. I like to own that function. In my role, I like to reach out to the candidate myself, get them pumped up about the opportunity, tell them what that entails, and then put them into the funnel and track them closely. Some of the best recruits I have made is by proactively reaching out to people. They were not looking. I had to convince them to join us on the mission.
So for these highly qualified recruits, what were the things that you identified in them that made you believe that they were the right candidates to be part of your team?
Both in Uber and Glovo, I restricted myself to people who have worked in the startup ecosystem. I’m not a stickler for past experience, as much as I am for passion and what the candidate brings to the table. I look at whether the candidate has had a good career trajectory over the last 2-4 years in the startup world, and if the candidate is willing to go the extra mile. This means – is the candidate really researching about us and reading everything that’s out there before interviewing? It is common etiquette when you are interviewing, but most people don’t do it. So if you’re doing this, you’ve already put yourself in the top 5%.
So I look for these qualities, and I want to see if they’re really pumped to work with us. Because at the end of the day when you’re expanding, when you are launching in a new country, it’s like a startup. You need passionate energetic problem solvers on the ground who can be completely autonomous to build up the business for you with minimal intervention.
A lot of our subscribers work in tech, and they would love some insights on how they should think about geographic expansion. I have a two-part question here. First, how do you know when you’re ready to expand geographically? Second, can you share a high-level playbook on how to blitzscale into new geographies?
I think geographic expansion starts when you have found product market fit in the initial home market. The definition of product market fit is different from one startup to another, so I would leave that up for interpretation. I, for one, come from an aggressive mindset when it comes to expansion. I’m a big fan of Chamath Palihapitiya. He was the early growth guy at Facebook and currently runs a VC firm called Social Capital in the Valley. He has this quote: “It’s f***ing land grab time, so get all the f**ing land that you can get.” This is his way of saying that if you have a product market fit, just go out and get as much of the world as you can. That’s what I believe in – once you have product market fit, you should expand as quickly as possible, also because your product can be easily replicated. Uber faced that with multiple competitors across the globe, and the same with Glovo. So it’s best to get in early into the market.
In terms of the playbook, people stress too much over the playbook. It’s different from startup to startup. Instead of a playbook, I like to refer to the major tenets crucial for hyper growth:
First is team. I think what Uber did extremely well in the early days is that they never compromised on the team. They paid that extra 10 – 20% to build the initial team. I fundamentally believe that A players bring A players, and B players bring C players. Uber had a cut-throat work culture: they would fire underperformers. It was a culture of really high performance.
Second are tools and processes. Uber had one of the strongest tech stacks that I had seen in 2014. Encourage data transparency. Pulling out data should be simple – not everyone is an SQL expert! Visualize the data, present it in an easy and digestible way for the teams across regions.
Employee training / on-boarding team – invest heavily into this right off the bat. Enough reading material, how tos, FAQs, etc.
Autonomy / Empowerment – If you hire an A team, have the tools / processes to empower them, and share best practices, let the team run as a decentralized unit. That’s the key with hyperlocal businesses that a lot of businesses get wrong – they tend to apply the same principles from their home market to new regions which doesn’t always translate well
Lastly, 996 culture – 9am to 9pm, 6 days a week. In a hyper-competitive world, if you aren’t working your ass off, and your competition is, they will definitely kick your ass. Jack Ma endorsed this recently and got a lot of heat for it.
Your comment about hyperlocal market, is that true for any B2C products, or are there any exceptions to that?
There are a lot of exceptions to this. Facebook and Google probably do not need any localization at first. When I say hyperlocalization, I am referring to logistics-heavy, on-demand businesses, which have a big workforce on the ground. And by workforce I do not mean full time employees. The local workforce could be partners on the ground: drivers, delivery riders, etc. Whenever you have a business like that, you have to have a very local mindset. Facebook doesn’t need to do this. They can just flip a switch to turn on a new market, and as the market grows, they can localize. So it’s different for different businesses.
As an investor, you must have seen a lot of different startups solving various problems. Any interesting problem space that you wish a startup would have been tackling right now?
It’s still early on in my investing journey. Something I’ve thought about is digitizing the process of applying & obtaining a visa. This comes from a personal pain point of having an Indian passport, which is not very strong for travel purposes. Hypothetically, with a certain profile or background, one shouldn’t have to go through the process of procuring a visa. I’ll gladly give read-only access to my tax returns, bank account statement, and other paperwork. Can’t they just automatically grant me a visa based on this information?
This is clearly a very hard problem to solve because it involves public policy and local government, but I believe it can be done. Obviously the incumbent here is VFS, which operates in multiple countries and is very profitable. Can you disrupt that and automate the process for a small majority – people with certain profiles, who have a certain background, and travel often? Why should we go through this archaic process when we are living in a technology-enabled global world. This idea actually came from reading about a startup from San Francisco called CLEAR, which expedites your security check-in at airports across the US.
Yeah – CLEAR is an interesting business because it actually competes with TSA pre-check (TSA precheck is a US government program where you can pay a fee of $85 for five years, enabling you to speed through a separate security line – where you do not need to remove your: shoes, laptops, liquids, belts and light jackets – for domestic flights).
To me, as someone who was on the road every second day for Uber and Glovo, I would pay for that kind of service. The amount of time one can save from this type of solution is significant.
‘Growth Hacking’ has become such a popular term in the startup world right now. What – according to you – does growth hacking actually mean, and how does it change from an early to a late stage startup?
Interesting question. To me, growth hacking at an early stage startup is about being scrappy with your marketing spend and being at your creative best. This means hyper experimentation across various channels of communication and distribution. So I’ll give you an example here, highlighting the different phases of the startup.
Back in the early days of Uber India – this is 2014 – whenever we launched Uber in a new city, in the initial few weeks, we would frequent the city bars, meet random people there and walk them through the app. This was when Uber wasn’t that well known in India. Why bars? Because taking an Uber to and from the bar is one of the most common initial use cases of the app. A few weeks in, we would strike a partnership with the same bar giving out free or discounted rides to all their patrons – and use the social media and network of the bar to reach out to a wider audience. And today, the same bar would use Ubers for themselves to get their regular customers home in order to provide that extra service and differentiate themselves. You can see the evolution of marketing here with one particular bar in a city.
Again, growth hacking will be different from startup to startup, depending on the vertical and the need of the hour – but principally the ethos behind the idea of growth hacking is doing more with less, and seeing indicators in your user lifecycle through the product for you to double down on. For example, Facebook: if you get to 10 friends within 7 days, your usage increases exponentially – and from thereon it is all about how does the platform recommend you friends as a new user to get to that threshold ASAP.
To close, I wouldn’t know much about growth hacking at a late stage startup as that’s not something I enjoy – however I’d think it’s about shelling out the big bucks for billboards, TV ads and signing brand ambassadors – something Uber has been doing over the last couple of years.
Follow Utsav: @utsavagarwal, or connect with him via LinkedIn