Monthly Archives: June 2015

SaaS metrics

Software as a service (SaaS) businesses shift a company’s fixed upfront software costs into variable pay-as-you-go costs that occur in the future. This makes SaaS businesses attractive for customers while also giving a relatively more predictable revenue base for the software provider.

However, when software is offered as a service, it’s no longer sufficient to evaluate the health of the business based on the traditional metrics of a license-based software provider. License providers care mainly about how many customers buy the license and how much revenue they earn from these customers.

In addition to these two core metrics, SaaS businesses need to monitor several more.

First, the shift to a pay-as-you-go model lowers long-term customer lock-in and gives the customer the chance to stop using the software each month. This makes it very important to monitor a SaaS business’ churn.

Second, customers’ ease of churning makes customer service very important. In the licensing model, customer servicing costs are largely incurred at the time of sale. However, in the SaaS model, these costs occur on a monthly basis and can eat in to your monthly recurring revenue (MRR).

Finally, the makeup of a SaaS business’ MRR is a strong indicator of the health of the business. The Social + Capital Partnership, investor in leading SaaS companies like Yammer, Box, and Slack, recommends breaking net MRR into new MRR (from new customers who begin using the software), expansion MRR (account upgrades from existing customers), contraction MRR (account downgrades from existing customers), and cancelled MRR (from existing customers who stop using the software). You can view their full presentation on what they look for in SaaS businesses here.

Especially at the early stages of a startup, it’s important for the company to be increasing its MRR by acquiring new customers rather than through account upgrades from existing customers. If expansion MRR is responsible for net MRR increasing for a startup where new MRR is otherwise outpaced by the contraction MRR and cancelled MRR, this should raise a red flag.

Social + Capital defines the quick ratio as the sum of new MRR and expansion MRR divided by the sum of contraction MRR and cancelled MRR, and looks for a quick ratio higher than 4 in its investments. Given its track record, this is probably a good target.


Postmates is the leading on-demand goods delivery startup in the US. Although restaurant deliveries which have a high transaction frequency account for most of the company’s volume, Postmates also delivers a wide range of convenience store products from its General Store as well as products from other non-restaurant stores like Apple and 7-Eleven.

The company is riding the wave of on-demand services made possible by a network of smartphone-enabled workers. There are large companies like Uber, Google, and Amazon competing in the space, but we believe that the ultimate winner will be a focused and nimble startup. The fact that Postmates has emerged as the strongest startup in the category is why we participated in its latest $80M funding round led by Tiger Global.

We believe that there are very strong network effects in the on-demand goods delivery space. Consumers want to be where the merchants are and vice versa. And having the most consumers and merchants on your platform allows you to fulfill the most orders. This liquidity makes your platform the most attractive for delivery workers and drives down their unit costs. Postmates has already shown that it can make money on each delivery and the unit economics will further improve as it scales.

Similar to the on-demand transport space, the significant network effects in the on-demand goods delivery space make it likely that there will be a runaway market leader within 3 to 5 years. Postmates is well positioned to fill this role.

Dropping out

I read a post entitled The Startup Illusion from Noah Benesch on TechCrunch yesterday. Noah is a junior in college and he shares how he recently considered dropping out of school to pursue his startup. He was about to do so when his team came apart and he decided to stay in school.

There’s so much hype around startups these days that many young college students, or even high school students, want to be the next big entrepreneur. There are also enough examples of very successful college dropout entrepreneurs like Bill Gates, Steve Jobs, and Mark Zuckerberg to justify experimenting. The problem is that for every successful dropout story you hear about, there are thousands who don’t make it. For every under 30 year old multi-millionaire you hear about, there are thousands at the same age who are left without a college degree after a failed startup. Sometimes they just weren’t cut out to be entrepreneurs, and sometimes market forces weren’t in their favor. In either case, it’s a tough spot to be in to rebuild your career.

I also worked on a startup while in college. It was a company called ExStat, short for Extreme Status, where users would create polls and rank everything in the world. Rather than officials ranking the best basketball team in the world, the masses would. The same would hold true for ranking colleges, musicians, and rankings in other fields.

I was a solo founder (not necessarily wrong but this stacks the odds against you), had an outsourced IT team build the website (definitely not recommended), and targeted all subjects that could be ranked from the outset (effectively not targeting any of them, because unless you have tremendous resources, it helps to focus on succeeding in a specific niche before expanding elsewhere). I also wasn’t aware how much reiteration after the initial release would be required to achieve product market fit. I mistakenly thought that users would simply come to version one. I launched ExStat on the MIT campus and a few polls were built and votes were given, but the site clearly didn’t take off. I was two months away from completing my classes so I gave up on the startup and graduated. We started Romulus Capital a few months later.

Every stage of ExStat’s journey was an opportunity for me to drop out of school. And although I didn’t, there are thousands of students each year who do. They drop out after forming their initial team, after getting some angel funding, after launching their product, or at another stage. But for most people, it doesn’t work out.

There’s no formula to tell you what the right decision is for your specific case. But given the hype around startups, more and more students are thinking about dropping out. You need to be honest about the specific reasons why you want to build a startup and what you’re willing to sacrifice to do so, rather than driven by the general hype around startups, to make the right decision for you. If you believe that your answers to these questions place you in the top 0.1% of students having the same debate, maybe you have a chance. Otherwise, finishing college is still a pretty good bet. It’s also what Elon Musk, Peter Thiel, and Reid Hoffman did.

Apple Watch use cases

It’s been a week since I started using the Apple Watch. As promised in my earlier post on the Watch, here’s a summary of the use cases that I have for it so far:

1. Ordering taxis: I find myself calling taxis from Bitaksi a lot more frequently on my Apple Watch than on my smartphone. The experience is equally fast once you enter the app on either platform, but I’ve found myself using the Watch as its presence on my wrist makes it more accessible than the smartphone in my pocket.

2. Messaging: I really like the short messages (OK, Thank you, Yes, No, …) and emoticons available on the Watch. I still use my smartphone for longer messages but the Watch does a great job of delivering to the point messages. While such short messages may be considered rude when delivered from a smartphone, they’re not when sent from the Watch. I think that the Watch is going to make to the point messaging a more acceptable social norm in the future.

3. Navigation: After entering a destination on my smartphone, it’s much easier to drive while glancing at the directions from my Watch than those on my smartphone. Using the Watch allows me to keep both hands on the wheel so it’s much safer.

4. Calendar: Adding new calendar entries remains the domain of the smartphone. However, it’s much easier to see my existing meetings by glancing at the Watch calendar than by pulling out my smartphone from my pocket.

The use cases I’ve highlighted show that the speed of communicating on the Watch, and the hands-free interactions it makes possible, are its two most important attributes for me.

The navigation and messaging use cases are in line with those that Bijan Sabet, a VC at Spark Capital, shared in his post about what he uses the Watch for. He didn’t see the same benefit that I saw from ordering taxis and checking my calendar on the Watch. However, Bijan also highlighted the benefit of the Watch for boarding planes and fitness tracking. I haven’t boarded a plane in the last week so I have yet to test this experience. And although I exercise regularly, I prefer to keep my exercises technology-free to avoid distractions.

Bringing together the many use cases that we’ve identified, I agree with Bijan’s assessment that “for a first generation product, that ain’t bad”.

Mobile deep linking

When you click on a link from one domain to another on the web, you expect the link to take you to the specific page within the domain that you just requested. For example, if you’re browsing a social network and you click on a specific news article, you’re taken directly to that article’s page rather than the homepage of the website.

Unfortunately this seamless experience that exists for web browsing is currently very rare for mobile apps. When you click on a link to a mobile app from your web browser, another mobile app, a text message, or another source, the link takes you to the download page of the app if you don’t already have it installed on your smartphone, or to the app’s homepage if you do. The specific page that you should be taken to within the app is lost in translation.

Wouldn’t it be better if the app could recognize what source you were coming from and take you to the right page within the app? If you were browsing a website and decided to continue your experience in an app, this would mean keeping track of where you left off on the website and starting your app experience on the same page. If you were referred to an app by a friend, this would mean personalizing your app experience by, for example, assigning you credit based on the source of your referral.

This is a practice called mobile deep linking and Branch Metrics, where we invested in their latest $15M round led by NEA, is doing exactly this. The company which launched in September 2014 already has over 500 developers at companies like Microsoft, Instacart, and HotelTonight using its service. Together these developers have created over 20M mobile deep links.

I believe that transitions to and from mobile apps will be as seamless as browsing the web in the near future, and that Branch Metrics is uniquely positioned to make this happen.

Employees and independent contractors

The California Labor Commission recently ruled that an Uber driver was an employee. Until now, Uber has treated its workers as independent contractors, thereby avoiding paying them employment taxes to cover benefits like unemployment, Social Security, and health insurance.

The ruling was made for a single employee in a single state, and Uber is appealing the decision. As a result, it’s not confirmed, and may not apply for the specific circumstances of other drivers and other states. However, for the first time, there is a real possibility that on-demand companies will need to reclassify some or all of their workers as employees. Estimates suggest that a full reclassification would increase their labor costs by around 30%. They could try to recoup some of this cost increase by raising the price they charge consumers and/or increasing their take rate but these would negatively impact demand and supply respectively.

A few days later, online grocery delivery service Instacart announced that it will be reclassifying its shoppers from contractors to part-time employees. Note that this is valid for Instacart’s shoppers and not their drivers who they will continue to treat as contractors. The reclassification is also taking place from contractor to part-time employee, not full-time employee. I don’t know if the timing of this announcement has anything to do with the Uber ruling, but it could be an effort to pre-empt a similar ruling for Instacart.

Beyond Uber and Instacart, the ultimate outcome of these rulings will impact other companies in the on-demand space like Lyft, Postmates, and Shyp. There is a lot at stake. The ultimate solution is very likely to be different from the status quo where these workers are treated as independent contractors. I think that there are two possibilities which stand out. The final outcome may also be a combination of the two.

The first option is that some workers are treated as employees whereas others remain contractors. For example, the workers that meet the base demand for a company’s offering (and thereby work exclusively for the company) may be treated as employees, and the workers that need to be brought onboard to meet peak demand (and thereby work for the company on a flexible basis) may continue to be treated as contractors.

The second option is for a new, third category of worker to emerge. This category would fall somewhere between an employee and an independent contractor. In discussions so far, they’ve been called dependent contractors. Dependent contractors would receive an hourly wage like independent contractors while they’re working for a company. However, their benefits like unemployment, Social Security, and health insurance would be unbundled from the companies they work for, and paid for by these companies on a pro-rata basis according to how much time the worker spends serving each company.

Technology in Istanbul

Fred Wilson is the co-founder of VC fund Union Square Ventures. The fund’s investments include Twitter, Tumblr, Foursquare, Etsy, and Kickstarter. His blog AVC, where he blogs daily, was also the inspiration for my daily blogging.

Fred and his wife spent their wedding anniversary in Turkey this weekend, and Fred blogged about technology in Istanbul yesterday. My key takeaway from the post was Fred’s correct assessment that Turkey is “… a place where social, mobile, local, messaging can take off as well as anywhere in the world …”.

In addition to Yemeksepeti’s recent $589M exit, our investments in companies like Bitaksi, Modacruz, and prove Fred’s point.

Valuing seed and early stage companies

I watched the interview below from the On Demand Conference yesterday. It’s between Semil Shah, founder of the Haystack Fund, and Shervin Pishevar of Sherpa Ventures.

In addition to being an investor in Munchery which I wrote about earlier, Shervin is also an investor in on-demand companies like Uber and Taskrabbit. I found the whole interview very valuable to see how one of the world’s best investors in on-demand startups thinks about his investments in the space.

However, the most interesting part of the talk for me was when Shervin shared how he led Uber’s Series B in 2011 while at Menlo Ventures. At the time, Uber had annual net revenue of $1.8M (a 20% commission on $9M of ride value) and Shervin valued the company at $290M. Shervin doesn’t state whether this was a pre-money or post-money valuation but it doesn’t really matter. Given the $37M round size, this implies a 140X to 160X revenue multiple.

Looking at the valuation from the perspective of a revenue multiple, it’s very difficult to justify. As Shervin shares, the reason he was willing to pay that much for the company was because he had a vision for how big it could get. This is a great reminder that, let alone a discounted cash flow analysis, even a multiples-based analysis doesn’t work when valuing seed and early stage companies. A much better approach is to think about how big a company can get, assign a probability to it reaching that outcome, and apply your target rate of return on the capital you’re investing.

The key determinant of how much you’re willing to pay is clearly how big you believe the company can get. Seeing this requires tremendous vision.


I recently received an invitation to join a mentorship program. Although I was initially tempted to accept the invitation, I had an uneasy feeling. I therefore decided to dig deeper and think about my past experiences as a mentor and as a mentee. In the end I declined the invitation. Here’s why.

I’ve been fortunate to have some great mentors who have guided me at different points in my life. Some of them continue to this very day. There are also two individuals to whom I believe I’ve been a valuable mentor, helping them make educational and career decisions that I made and learned from in the past. So I certainly believe in the value of mentorship.

However, each of these mentor-mentee relationships developed serendipitously. They arose from a personal bond based on initial shared interests and the subsequent discovery of shared values as we got to know each other better.

This informal development is in contrast to formal mentorship programs where mentees and mentors are matched by a third party through an organized application and review process. I’ve participated as a mentor and as a mentee in such programs, and in each case the relationship faded away after 2 to 3 meetings.

Based on these experiences, I believe that the best mentor-mentee relationships develop naturally. They’re sparked by a shared journey that two people are embarking on at different times, and develop into a deep bond if these two people also like spending time with each other. Trying to impose a structure on this natural development lowers the quality of the matches made and the value that both the mentee and the mentor get from the relationship.


Digital signatures

We invest in both Turkey and the US. Whenever one of our US startups has legal documents that need to be signed, we simply do so online. US law accepts digital signatures and this makes signing legal documents very easy. We simply go onto Docusign, Hellosign, or whichever platform the startup is using and sign the documents there.

The process is much more cumbersome when one of our Turkish startups has legal documents that need to be signed. Turkish law requires that legal documents be signed by pen and paper. In practice, this means that the startup’s lawyer takes a physical copy of the documents to the location of each party that needs to sign it. They travel from office to office in Istanbul, obtaining successive signatures until all parties have signed the document. This is a time sink for the law office and it creates an additional cost that the startup needs to pay for. It also requires each party to coordinate their schedules to be available when the documents will be arriving at their office. When delays occur, this impacts everyone’s schedule.

The requirement for pen and paper signatures is archaic. We live in a digital world. Turkish law, and likely many other legal systems across the world, need to start accepting digital signatures.