Nice job - congrats. Is it difficult to prevent gaming the results? I'm curious about "pump and dump" approaches that exploit the social media data stream.
It's an interesting question and we haven't had any problems so far. The good thing is that our data feeds are vast in terms of both size and reach across the social web, and we don't overly rely on any one feed.
Just hackin' away with PHP, jquery, and MySQL ^^ The graphing libraries can be found here http://www.highcharts.com/ They've been uber easy to integrate
haha.. Things are starting to get a little crazy in that area, but for now we are getting away with a few synced up designated data crunchers (crawling and crunching in C++ goes a long way as well!)
Hahaha, way to take the Recurly form and use it with Stripe. I did the same thing. I'm not too familiar with following the stock market, but good luck to you guys and the product looks great!
It seems like you would make more money keeping this technology secret and using it yourself or convincing a single investor to use it and share the proceeds. Why are you trying to market it?
haha we are definitely trading on it ourselves, but as entrepreneurs we don't have nearly enough capital to make the profits meaningful. Also we're really excited to be making affordable information tools for independent traders and have a lot more in the pipeline! :)
If everyone knows the secret it is not a secret anymore and the information becomes useless. If too many people are trading using the same algorithm the profit will be squeezed out and the data will be about worthless.
Find a Hedge Fund manager somewhere and get some money to put into the strategy. Of course you may find then that the simple act of investing larger sums will destroy your edge as your trades start having more significant market impact.
We're not peddling a stock market predictor, it's an information tool to give you an idea of how companies you're interested in are performing in real-time to help you make long-term decisions. We're not doing high frequency algo trading.
You have a tool that quantifies real-time performance of indivicual stocks. This information is helpful to make long-term buy/sell decisions. Making such decisions based on this information is apparently profitable.
Unless they delay the publication a bit, and profit on the volume of trades done by their subscribers. The old dirty trick of moving prices of low-volume stocks.
Unrelated, but part of a bunch of work my startup is doing is creating a set of new data sources around media coverage of big listed companies.
At the moment, we produce all the data retrospectively, but we could do it in real-time(ish). We think there's a strong chance that much of this information is tradeable, but we're very very busy doing the work that we sell to customers.
Anyone got any good ideas on how we go shopping for a customer for this data, given that we have no time? For some clients, we have many many data sources that aren't available publically(ish - we create the data from publically available data) going back several years for the marketing departments of some huge clients.
Presumably we need a quant who we trust with some time on their hands to play with it...
A quant will definitely be important, especially one with a background in finance. However, you said the data wasn't available publicly, this could be a huge problem if you are planning on selling to the financial sector.
as an investor, i do wonder "if its working well consistently, why wouldn't they keep it a secret?" So that makes me immediately suspicious as to why I'd want to use it if it wasn't working well enough for you to use yourself.
Just curious how you answer that question b/c if returns are good enough, you should be able to raise money for a fund w/ no problem.
It's definitely a great question!
The answer is simply that we didn't have enough investment capital for it to be meaningful, though we do still trade on it to build our track record to show investors. All of our angel investors came on board only after first using the product. But also we feel like we have priced it so that people can try it out fairly painlessly and decide for themselves if they find it valuable. Though we haven't had any unsubscribers to-date ;)
So, can we please separate facts from fiction? The Forbes article states that TTAGG's back-end downloads all photos that users post on their Twitter accounts and then tries to recognize whether the users are wearing what they claimed to have purchased in their Tweets. Is this really the case? How can you tell apart a pair of American Eagle jeans from a pair of Levi's?
that starts to get to the core of our technology, so we're not really comfortable discussing it. That being said, in your example there are ways to derive the associated brand based on information alongside or attached to the photo.
Well now, that's an assurance that it wouldn't work, but not an explanation of how.
I understand you don't want to give away all your secrets, but any hints? Any assurance that you really have worried about the problem that someone might try to manipulate your system?
If this data is as actionable as it seems don't piecemeal subscriptions, you will rob yourself long term by dealing with all the headaches of direct-to-consumer(ish) selling.
Revisions and accuracy on the data set and protecting that IP is where you're going to extract maximum value.
A good place to start calling would be Bloomberg - they have dialed in how to extract maximum profit out of business data.
They have a venture division that would be a perfect fit http://bloombergventures.com and can push you in the right direction if they pass.
Worth checking out this article on StormPulse, they were in a similar position - incredible data - just needed to get their head around monetizing it properly:
I work for a quantitative stock trading firm, and if this data is an effective predictor of stock returns you are underpricing it by a factor of about 4x to 25x. Most data providers charge at least $10,000/year, with the more effective ones charging in the six figures.
The primary thing a quantitative trader would want to see is a rigorous track record. Ideally this would be a "point in time" record where you record the information that would have been available as of a certain time in the past. If you don't have the data recorded for that, or it's for too short a time period, the next best thing would be to rerun your metric on all the historical tweets you can get your hands on.
Showing potential customers a mostly complete historical dataset of what you would have told them is a fairly common way of selling this sort of data. Don't show them anything too recent so that they need to buy the data from you. Firms have different investing processes, and so they often need to see a larger sample of data in order to know whether it fits well with how they invest.
Please feel free to contact me if you'd like to talk more about this -- my email is my handle at Google's mail service.
40 comments
[ 3.2 ms ] story [ 94.0 ms ] threadBut From the wilds of Fayetville, Arkansas? Please.
This looks like it might be disruptive, and in particular to an area that needs disruption.
Find a Hedge Fund manager somewhere and get some money to put into the strategy. Of course you may find then that the simple act of investing larger sums will destroy your edge as your trades start having more significant market impact.
How is that NOT a stock market predictor?
At the moment, we produce all the data retrospectively, but we could do it in real-time(ish). We think there's a strong chance that much of this information is tradeable, but we're very very busy doing the work that we sell to customers.
Anyone got any good ideas on how we go shopping for a customer for this data, given that we have no time? For some clients, we have many many data sources that aren't available publically(ish - we create the data from publically available data) going back several years for the marketing departments of some huge clients.
Presumably we need a quant who we trust with some time on their hands to play with it...
Just curious how you answer that question b/c if returns are good enough, you should be able to raise money for a fund w/ no problem.
Step 1: I go long on Nike
Step 2: I use my ten thousand twitter sock puppets to tweet about how they just bought an awesome new pair of Nike shoes
Step 3: I watch as all the TTAGG users plunge money into Nike.
Step 4: Profit!
I understand you don't want to give away all your secrets, but any hints? Any assurance that you really have worried about the problem that someone might try to manipulate your system?
Revisions and accuracy on the data set and protecting that IP is where you're going to extract maximum value.
A good place to start calling would be Bloomberg - they have dialed in how to extract maximum profit out of business data.
They have a venture division that would be a perfect fit http://bloombergventures.com and can push you in the right direction if they pass.
http://gigaom.com/2011/09/20/how-stormpulse-made-more-money-...
The primary thing a quantitative trader would want to see is a rigorous track record. Ideally this would be a "point in time" record where you record the information that would have been available as of a certain time in the past. If you don't have the data recorded for that, or it's for too short a time period, the next best thing would be to rerun your metric on all the historical tweets you can get your hands on.
Showing potential customers a mostly complete historical dataset of what you would have told them is a fairly common way of selling this sort of data. Don't show them anything too recent so that they need to buy the data from you. Firms have different investing processes, and so they often need to see a larger sample of data in order to know whether it fits well with how they invest.
Please feel free to contact me if you'd like to talk more about this -- my email is my handle at Google's mail service.