Monitor110: flatline. But…
Today Silicon Alley Insider reports that Monitor110 has closed down:
The company had raised $20 million, including an $11 million round in 2006. Last month, we reported that the company had cut a third of its staff and was trying to sell itself to Reuters. Carley subsequently told The Deal that he was looking at several options: “So while we have been looking for investment and distribution, we are open (given the right terms) to an outright sale of the company as well, and several parties have expressed that kind of interest.”
From the start, back in September 2006, when they burst on to the scene with much fanfare, while I wholeheartedly agreed with the concept (of helping traders vastly broaden their information horizons by tapping into the exponentially growing alternative sources emerging on the web,) I was politely sceptical as to whether or not Monitor110 had the right approach. To be clear, beyond what was available on their website, I didn’t know much about their technical approach or even get to see what their product looked like. So that’s not what put me off. Rather, what made me uneasy was that they seemed to be taking a very traditional Wall Street approach - business model, budgeting ($20mn !!!???), even look & feel - to what for me was clearly a non-traditional proposition. It may be cliched but successful solutions in this space will have much more of a Web2.0 feel about them. (I know this is oversimplifying.) One company that I think may be on the right track for example is Skygrid. But there are many others. And given the cutting-edge nature of this space, even good ideas well executed, with really smart people like Roger might not succeed. At least not on the first try.
I guess the point is that opportunity abounds in this space and I would not be put off by Monitor110’s demise. I would however take away the lesson that selling information services using the traditional model (a la Thompson Reuters) is very hard unless you are the incumbent or you have a data-set that is immediately and obviously unique and valuable. Start-up companies in the emerging unstructured information/AI space have neither and so need to take a more creative approach to their business model and a more cautious approach to their burn-rate. Artificial intelligence, semantic search, natural language processing, trust(reputation)-based filtering, etc. etc. all will continue to grow in importance and relevance to all the various players in the financial services industry (and for that matter any industry where finding, organizing and synthesizing information is core to value creation.) Indeed this is very much the context for my interest in Powerset for example.
So I hope other budding entrepreneurs and technologists looking at this space will not be put off by this. And I hope to have invested in one or two of them by this time next year!
Update:
If you haven’t seen it already, please read Roger’s excellent post-mortem analysis of Monitor110. There are a number of lessons that are universal (to start-up ventures) - two key factors being money (yes, you can have too much) and leadership (clarity clarity clarity)… to give you a taste, here are his “Seven Deadly Sins” (for a start-up):
1. The lack of a single, “the buck stops here” leader until too late in the game
2. No separation between the technology organization and the product organization
3. Too much PR, too early
4. Too much money
5. Not close enough to the customer
6. Slow to adapt to market reality
7. Disagreement on strategy both within the Company and with the Board
July 18th, 2008 at 12:39 am
I completely agree about that some of these companies are or were pursuing an antiquated business model. And it’s also quite amusing to see these web services assume, in deference to Bloomberg, that they are more likely to succeed with a black background
.
But I think the root of the problem for many startups is that they are technology-driven rather than needs-driven. This is just as much true of Web 2.0 companies as it was for Web 1.0. In the financial domain, for example, you need to understand the pain points for portfolio managers, traders, analysts, and their teams. And then create innovative solutions to the specific problems. The solutions may blend UI techniques, reference databases, unstructured text search, and even some NLP techniques; what is important is the problem being solved.
This is the approach that companies like Tamale Software and ourselves at Atomic Intelligence have taken.
July 18th, 2008 at 7:01 pm
Sean,
I think that you are right about the approach and business model. Some of the key learnings from Monitor110 should be:
- It is very hard to deliver alpha-generating services from internet information, and in any event hedge funds and portfolio managers won’t outsource alpha generation.
- It is very hard to extract signals from the *entire* internet, which is inherently noisy.
- Which is why we re-focused the company on taking a hand-selected list of sources, extracting relevant content from those in real-time, categorizing that content by company and by investment theme (e.g. macro topics that are typically catalysts for movement in a stock, such as mergers, supply-chain issues, regulatory, technology change, etc.) and delivering those to institutional investors. This is generally not “tradable” information, and so the target market is not traders (sexy as that might sound) but portfolio managers and research analysts. The value to them is access to information that is well categorized and intelligently filtered, which saves them time and allows them to efficiently broaden their research coverage.
- While that is a valuable thing to do, it is not valuable enough to compete for the scarce money and even more scare attention from customers that are required to sustain a standalone business.
- In the end, I believe that the technology/product we built has value, but it has maximum value when integrated into existing distribution platforms (on on end this could be traditional platforms like Thomson Reuters or Bloomberg, on the other end this could be advertising supported distribution like Google.)
Unfortunately we simply ran out of runway before we were able to put these learnings into effect.