Ellen DaSilva's Profile

Ok Utsav, you’re breaking my heart a little here.

For full disclaimer, I worked for Twitter for 3.5 years on the business operations side and am fiercely loyal to both the product and the company. I’ll point out that I am writing on behalf of myself and that these views expressed are my own, not in any way reflective of the company’s. I think you have some fair points, but you’re at times I fear you’re missing the forest for the trees. I won’t have time to unpack every single one of your arguments, but here are a few high-level thoughts. Happy to answer in more depth in-person:

1. MAUs as a metric – monthly active users is a somewhat arbitrary way of measuring user growth. While I agree that it makes it easy to compare to companies like Facebook (which I would argue at this point is not so much a competitor as a complementary service), it disregards one of the major use-cases for Twitter, which is logged out. Twitter has a vast network of people who access Tweets on a very frequent basis. If you’ve ever read an article on Buzzfeed, you probably count among them. That number is many times the number of “logged in users.” And while that logged out userbase used to be less valuable because of the data and information available to the company about you, that’s no longer true. As of this week, Twitter announced its intention to monetize both logged in and logged out users. That is a huge opportunity for the company from a usage, data generation and monetization perspective that I believe will add huge value to citizens of the internet.

2. Developer stack – as you point out, Twitter had an acquisitive streak in which they bought companies like MoPub, Crashlytics, Gnip and a few others. They did this great thing by putting them all together into what’s called an SDK – a software developer kit, that enables developers to build using Twitter’s services from a data, monetization and testing perspective. This neat little kit has appealed to hundreds of thousands of developers, and in fact I believe Crashlytics is one of the most popular ways app developers test and develop code. This means that Twitter as a company is able to spread its wings far beyond the Twitter.com website and Twitter app. I would say this is akin to “login with Facebook,” although much more back-end heavy.

3. MoPub, the great behemoth – MoPub, a mobile ad server and RTB platform. Though MoPub’s syndicated web of mobile applications, advertisers on Twitter are able to access many times the number of MAUs on the direct Twitter platform. I’m hesitant to substantiate with numbers because I’m not sure what’s publicly available, but this is a huge opportunity for the Twitter monetization engine to spread its wings far beyond the walls of Twitter’s O&O products.

+1….great report Ian! In major cities where GRUB has achieved scale, I think their offering was for a long time revolutionary and is still a great service, as you describe. I might be a little biased because I wrote my TOM challenge on the competitor that’s seeking to disrupt GRUB :), but here are some thoughts to build on to Vesal’s:

GrubHub/Seamless has 5MM MAUs, but they are unable to penetrate many markets where restaurant-supported takeout isn’t as popular (for example, San Francisco). As Ves mentioned, they rely on a network of already existing couriers and delivery people, employed by the restaurants themselves. Where they fall short is when they haven’t achieved scale in a certain metropolis – for example: whereas in New York, there are many options to choose from, in smaller cities without that option, competition does not drive efficiency for the end user. A lack of competition means restaurants are more complacent with their services.

Therefore, as a many-times frustrated user of Seamless, I find myself blaming Seamless when something goes wrong, when really they are just the middle man. Given the way they’ve scaled, GRUB might find themselves in a tough place because they cannot control many of the operational aspects, but do bear most of the revenue/business risk when things go awry. I worry that as they expand to new markets, the costs will begin to outweigh the benefits if they cannot find a way to control the entire ecosystem. Their business and operating models might start to diverge, at which point it would be time to rethink the strategy.

Furthermore – I’m intrigued to understand why they haven’t begun to explore delivery options for other goods (supplies, alcohol, packages, etc). Is food the only thing people want delivered these days? Without the ability to control the operational supply chain, they might be limited in their ability to deliver other items. As we begin to shift toward the “on-demand” economy, I worry about how will GrubHub broaden its offerings to compete with peers.

I have been a member of OneMedical for 3 years and I’m a big fan. One of their draws is that you know what service you’re getting from city to city, so as I’ve moved, I feel comfortable knowing that OneMedical has already vetted my doctor and provides the same healthcare services. A few things come to mind as potential obstacles for OneMedical as they try to expand offerings to keep up with growing demand:

1. How large is their ultimate market? Are they only looking to expand within the US, or is this scalable outside of the states? Given different healthcare systems, I suspect it would be very challenging to move beyond the United States, but that also limits their business prospects.

2. How do they plan to compete with services like Pager and Doctors On Demand, that send trained physicians to your home to do the same thing. It would be a logistically tough thing to operate, but they might face competition from services like that, even for primary care.

3. Would they ever plan to tier pricing to cater to people with different needs (or packages for multi-city use), etc. The membership fee has gone up every year since I’ve used the service and while I find it valuable, for most people who access doctors once per year, the $200 fee is relatively steep.

Just a few thoughts – I really liked reading about how they optimize patient throughput through organized systems and teleconferencing!

This is fascinating, Will. I wonder if this vicious downward spiral of Amtrak’s is emblematic of broader infrastructure issues we are facing in the US. Is it possible that people’s travel preferences are changing, and there simply is no need for connected rail travel? I have to wonder if the fact that they make most of their money off of the Northeast Corridor route is the market’s way of telling them it’s time to shut down. If they did close, what would they do with all of the spare parts?

Alternatively, could Amtrak possibly pivot to focus their business model on something else? Similar to what Sam suggests, could they change their pricing structure to be more appealing to customers on less desirable routes? Could they partner with trip organizers to make Amtrak part of a whole package instead of a means to an end?

This is very interesting, thanks for elucidating something I usually just walk past without thinking. Given the nature of the goods and the travel involved, I’m wondering how they think about inventory. Do they target highly local goods so that each Dufry is unique to the market of the airport it’s sitting in? Or do they optimize for global goods that can be stocked anywhere? I imagine one has an impact on the business model, and the other, the operating model.

This is a great report, Sam! I particularly enjoyed reading about their consolidation of factories closer to their headquarters rather than outsourcing production to many different locations. Do you know if demand suffered as a result from the decreased number of pieces? My understanding with LEGOs is that even though you buy them in a set, pieces are meant to be used generally and creatively. With a limited number of pieces available (down from 12,000), did consumers feel that their ability to produce was limited?

In addition, do you know why LEGO opted not to bring all operations in house and vertically integrate? For example, would it be beneficial for their economies of scale to own the entire supply chain rather than just the factories?