Foursquare honestly has some of the best user location data you can buy on the market.
The only companies out there with better historical data are Facebook, Apple, and Google. And as far as I know, I can't call them up and start buying from them within a few weeks.
I've seen hedge funds right now making millions off of Foursquare's data as they can build algorithms around it quite easily and there's plenty of history to backtest off of. The great part is that Foursquare's previous money maker was selling their places database so they've already done a great job mapping GPS coordinates to businesses and it's pretty easy to map that further to a ticker symbol. Perfect for quants.
By combining Foursquare data with anonymized credit card data from Intuit or Yodlee... you've got the ability to predict retail sales on a DAILY basis rather than on a quarterly basis. You might even be able to a better job than that actual business in predicting their growth if you've got a whole team of quants working on creating the data models.
This is the start of a shift from the market at large relying on quarterly earnings for these types of companies and instead being able to track performance on a daily level. As far as I can tell, there's nothing companies can do to stop this either.
How do they continue to get accurate data about location? Fewer and fewer people "check in" to foursquare now. Presumably thousands have uninstalled the app. How will 4sq solve this going forward?
They still get millions of location updates every 15 minutes. I don't remember the exact number as it has been a couple years since I talked with people using their data but it is much much higher than you would expect.
They don't need you to "check-in" when they are grabbing your location every 15 minutes in the background. They've also built APIs into ad networks that ping their location database with coordinates millions of times a day to find relevant ads. There's thousands of apps using ad networks that funnel location data back to 4sq.
Techies may have abandoned them, but they still have a ton of middle America thinking that 4sq is cool. My mom for example.
On another note, 3-4 years ago I had access to a large anonymized bank/credit card transaction database from a well known company. It had about 1-2% of all transactions in the US by our estimates. When we modeled the quarterly sales of Walmart compared to it we found that it was incredibly accurate going back 5 years, even with 1/50th of the US population.
> On another note, 3-4 years ago I had access to a large anonymized bank/credit card transaction database from a well known company. It had about 1-2% of all transactions in the US by our estimates. When we modeled the quarterly sales of Walmart compared to it we found that it was incredibly accurate going back 5 years, even with 1/50th of the US population
I used to work for an Australian consultancy "Quantium" who are doing exactly this and scale transaction data using national statistics data [1] - page 13
Some Capital One employees had the same thought and hijacked their data to make $4.4M over ~15 months:
> "The SEC sued colleagues Nan Huang and Bonan Huang in January 2015 alleging they made hundreds, if not thousands, of keyword searches on their company's private database for sales data on at least 170 publicly traded companies from November 2013 to January 2015."
It seems like the biggest problem was misusing non-public Capital One data. Imagine if a personal finance app made that part of their business model and got scale.
Something like Mint could front a powerful consumer investment strategy..
There's a great article on it that I'm struggling to find. But essentially, the Shazam app and related apps built on their tech are used as a massive A/B testing tool now by artists. Roll out 3 versions of the same song to 3 similar cities in the Midwest and see where it gets the most shazam lookups.
If I recall correctly a few years back they were doing $100m a year by basically being able to give everyone in the music industry insight into what's the most popular and upcoming songs in every major city.
However, with the popularity of Spotify I'm guessing they're losing a lot of market share in the mass music metrics space.
> They don't need you to "check-in" when they are grabbing your location every 15 minutes in the background.
If they are tracking users who are not actively using the app then that is just plain creepy and wrong. Are you saying Foursquare tracks background location to thus day for users who installed it and tried it out once a few years ago? That would explain the database value but it's Creep City.
Operating systems should expire those background privileges for apps that fall into disuse (or do they already?)
On the iPhone you can go to "location services" and see a list of who's tracking you. I just did and there are 8 that say "always". You can turn it on and off but few people including myself bother much.
Foursquare has been pretty much unused in The Netherlands, as it is a small but important market as they are rich to the max. If your service doesn't succeed or sell in The Netherlands, it's usually indicative to your service at large.
For the Foursquare/Swarm apps, when they are able to track your location 24/7, they create a list of places that you may have visited and allow you to create an after-the-fact check--in. That's what's obviating the need for users to actively check in to the app(s).
They likely have multiple streams for data ingestion including Swarm/Foursquare and the signal data provided by thousands of developers that use their API including some large folks like Uber, Apple, Snap, Samsung, Twitter, Microsoft etc.
Even if both of their apps fail, they likely have a lot more inbound data from these large partner sources to continue to build out a pretty healthy pipeline to provide better data.
I'd estimate about 2% of my Facebook friends are regularly using Swarm, and about half of them very heavily (multiple checkins every day). It would be foolish to extrapolate from a sample of 1, but I find it likely that they still have tens of millions of active users. Maybe they have something in the pipeline to make Swarm more interesting?
I was a bit surprised that the article didn't include 'Foursquare Hedge Fund' as yet another way to make money. Seems like they could exploit their own data just as easily.
Of course I really have no idea how difficult it would be to 'start' or manage a hedge fund. However I have met people who have run hedge funds and presumably such a person could be hired. Or more simply an exclusive partnership could be established with an existing hedge fund which was already able to use the data.
The underpinning of my surprise was the business case analysis. Your the CEO of Foursquare and you're looking at all the ways you can use your assets and services to generate revenue. You come up with selling this sort of demographic data. That gives you revenue $X. Now one of the buyers of your demographic data is hedge funds that are making a return Y % on their investments. Which increases a marginal amount dZ % when they incorporate your data into their strategies and algorithms. If you could accurately characterize the impact you could compare 1/2dZ % annually to the $X value to see which was larger. Conversely you could take say $10M and put it into the hedge fund and have them remit back to you 1/2 of the annual return each year and compare that to $X.
And of course if you're Goldman Sachs does it make sense to buy FourSquare and keep all the data for yourself? One could look at that investment in a similar way using internal rate of return on the buyout price vs the improved rates of return on their investment products.
Aside from the fact that they are a data company, not a fund... While what they are doing is completely legal it's safer for the company investing off of data to be separate from the company creating the data. Less conflicts of interest in the event it eventually becomes a legal issue.
No. Insider information has to ultimately come from an insider. Foursquare isn't getting its information from companies, it is getting them from its users.
Head of engineering at Foursquare here. Thanks for all the enthusiasm around our story. I’ve been at Foursquare since 2010 and can share some additional insight.
If you're an engineer building out a location-based product, and you’re curious how you can play around with our location data, we do have one of the world's most robust place databases. Twitter, Microsoft, Samsung, Snapchat, Apple, Uber, Pinterest —some of the world’s biggest tech companies— are building products on top of our Places database and API. We have a good-size free tier, and it's easy to get started: https://developer.foursquare.com/
The fact that it’s considered to be one of the most accurate databases, and that it’s widely trusted (used by more than 100K developers), didn’t come easily. The accuracy of our our place search API has been steadily improving over the past seven years. To do it well is a really challenging engineering problem. The sensor signals available from iOS/Android location services in today’s phones are often only accurate to tens of meters—especially when indoors. And this is an improvement! So we've taken the billions of check-ins we've collected since 2009 to basically create a model of what physical places look like from the perspective of a mobile device.
We continue to capture training label data from Foursquare City Guide and Foursquare Swarm usage, which allows us to continuously improve our place search model. A lot of complex machine learning and infrastructure goes into solving this problem. And by the way, Place search is just one of Foursquare’s many engineering challenges. We've been innovating and building new products over the past eight years. In March, we launched our Pilgrim SDK. There are even bigger challenges and opportunities ahead...
I know this because I've been here almost since the beginning: Right now is the most exciting time in the company's evolution.
It's important to mention that we take our user data privacy extremely seriously. Some of the products we're currently working on, including the ones that are mentioned by the Entrepreneur article, are built on anonymized and aggregated visit data. The products that we're building don't expose any individual visitation data at any point. We are forthcoming with our community about how we use our data, and we work hard to create engaging apps that make cities easier to use through search & discovery and checking in.
As we continue to rapidly scale and grow we are hiring in both SF and NYC for almost every area of engineering. We're especially interested in experienced iOS, Android, data pipeline (kafka, hadoop, spark, scalding), and ML engineers.
Hi. Since you brought up privacy.. Where exactly does Foursquare gather its location data from for these enterprise reports? Is it exclusively from "checkins" on Foursquare apps, or does it include locations provided via lookups to the Foursquare API places database from various 3rd party apps?
I've always wondered how deep the data could be if it was just Foursquare checkins, given the decline in usage of the apps. Yet I only ever hear about "billions of checkins" as the source. Not billions of queries. It doesn't quite add up for me.
I kind of feel that if Foursquare is tracking user locations via apps that use its API, that should be more clearly stated. Is it? Thanks.
i'm curious - how would you invest in order to take advantage of daily or presumably, hourly trends? what can actually close the loop that quickly on human activity -> financial securities performance?
ex. Goldman Sachs predicts Apple will have $25m sales Q2 2017 at the beginning and some variance of the stock price is based on that public prediction.
If you have finer granularity like Apple is likely already at 10m only a quarter through Q2 then you can invest and expect the Q2 prediction to be exceeded. Slice thinner and thinner until you get to hourly as the market adjusts to finer granularity predictions from companies buying Foursquare data.
For instance, if a company starts running an ad campaign that is particularly effective it will beat expectations in the next quarter. The more granular data you have, the earlier you can pick up the effect and buy the stock.
Apparently I've angered a handle full of people that don't agree with the statement above. Hypothesis: They're probably part of the group selling user info.
What this article tells me is to not share data with foursquare since their priority is to sell that data and make money and not to help me make decisions on where should I be eating.
Well to get you to share your data, they need to provide value to you which is helping you make decisions on where to eat... that's the web business model, you give up data in exchange for something, company sells data (and sometimes access to you too, like Facebook).
I don't see how you learning this changes anything. Either they were useful to you before or they weren't. What does it matter how high on their priority list it is? If you were their number one priority but they weren't helpful would you use them?
They're using your aggregate data to reward companies that you like and punish companies you dislike. It will make companies react to peoples preferences faster and provide a better product.
Of they would just track me, without any other features, it would still be a win.
But that priority doesn't influence the quality of the recommendations they make. The data they sell doesn't influence the places they tell you to visit because that would in effect bias the quality of the data.
> Foursquare is on the path to $100 million in revenue, and profits are within sight for the first time.
Serious question: how is a software company that's bringing in $100MM a year without any substantial overhead in terms of physical manufacturing or shipping or similar real-world expenses not profitable. As far as I can tell, they'd have to have infrastructure costs well into the double digit millions yearly and/or 450+ employees on payroll, neither of which is true as far as I can tell (regardless of how prudent either would be for them).
Let's do some quick napkin math. Foursquare has 286 employees[1], let's say the average salary is $120k, representing ~$180k fully-loaded cost. And there you already have a ~$51MM cost.
Probably a sales- and engineering- heavy team; many of whom will be earning $150k+. Executives are probably making $300k+. This should bring up the average salary at the company.
Also, this is just one line-item. There's a lot more expenses to running a business. I wouldn't be surprised to see them close to $80M+ a year. The article didn't exactly say their revenues were $100M, just that it was "well on its way" there.
The article says "Foursquare is on the path to $100 million in revenue", so they might be nowhere nearly that level. But I would think their largest costs are salary and servers.
Like most startups building a mass of users (e.g. Twitter), they have likely been optimizing for growth as quickly as possible. Which means spending all revenue/investment on getting customers, etc. and delaying focus on turning a profit until some future time (e.g., as the acceleration of growth slows).
"Looming large in this sector are Facebook and Google, each with its own armies of engineers working to compete. Glueck shrugs off these threats: The market is plenty big."
What makes Glueck think Facebook and Google wants to share? They enter big markets then attempt to dominate it, especially Facebook. Currently Facebook gives no quarter. It will copy what you do well and go after your business fiercely.
“Foursquare has signed deals with Snapchat to improve its geo-filtering. More than one million users have agreed to leave location sharing on all the time so Foursquare can track and analyze their movements”
Facebook(Instagram) and Google have significantly more people using their apps on location, the have more location data. Both companies do not have to buy data/service from Snapchat. And what happens if Snapchat and Twitter decides to enter the market?
That's pretty interesting, I thought they were dead for sure after "doing a digg" by splitting off Swarm from the main app. How do they even collect data if hardly anybody uses their app anymore?
I understand differently. Foursquare is not buying anything from Snap, they are selling their map technology. Pinterest and other companies are buying as well.
Thus, they sell a service and receive money and data from the users of those applications. I read other articles before telling how the information in their maps and APIs are impressive and the fact that you could use for free, if you are a small company, is a great deal. But, they are still not as valuable as they were before and they greatly depend on large companies. It makes me think that they are a great target to be acquired before becoming a unicorn.
It is quite clear that snap will bet heavily on AR and location. They have launched Spectacles and now a new way to discover content using location options. How long will they pay Foursquare to profit from their own data? Why do they not build a proprietary location technology or acquire another company that has?
Interesting, but they are essentially building a profitable business on the backs of people willing to be tracked (or checkin). You wonder how that can be maintained.
Aren't they getting a large portion of their data from 3rd party apps connecting to their api? Interesting products (like Pinterest and Snapchat) should make the steady inflow of data a safer bet.
Additionally, the user gets to tell Foursquare when to track my current location. Compare that to something like Uber, which is trying to always track you. I am more okay with this type of discretionary tracking.
I don't. I didn't read the article, I was bored and I thought parent was being sarcastic saying "building a profitable business on the backs of people willing to be tracked". Missing the key word "willing", I assumed it to be the same business model of just about every internet company and Foursquare was just now figuring this out. So I broke the rules and responded sarcastically thinking this post wasn't going to get much attention anyway (it's about Foursquare).
I've read it now and my comment makes little sense in context. I see they stated some users "have agreed to leave location sharing on all the time so Foursquare can track and analyze their movements." I'm not really sure how this is any different than what they were doing 8 years ago so I do, non-sarcastically, "wonder how that can be maintained".
I wish there's was a Zillow for businesses. I'm trying to buy a small business but you never know if they're lying and sales are dying. And I'd buy it then go out of business.
The only companies out there with better historical data are Facebook, Apple, and Google. And as far as I know, I can't call them up and start buying from them within a few weeks.
I've seen hedge funds right now making millions off of Foursquare's data as they can build algorithms around it quite easily and there's plenty of history to backtest off of. The great part is that Foursquare's previous money maker was selling their places database so they've already done a great job mapping GPS coordinates to businesses and it's pretty easy to map that further to a ticker symbol. Perfect for quants.
By combining Foursquare data with anonymized credit card data from Intuit or Yodlee... you've got the ability to predict retail sales on a DAILY basis rather than on a quarterly basis. You might even be able to a better job than that actual business in predicting their growth if you've got a whole team of quants working on creating the data models.
This is the start of a shift from the market at large relying on quarterly earnings for these types of companies and instead being able to track performance on a daily level. As far as I can tell, there's nothing companies can do to stop this either.