Monitor110 Learnings: The Good, The Bad, and The Really Bad
The response to my Monitor110 Post Mortem was pretty shocking. I want to thank those who commented, sent emails, and simply took the time to read the post. If some of you are really able to take my experiences to heart and to avoid some of the mistakes made by myself and the team, that would be terrific. To that end, I am going to spend some more calories drilling down on key areas of note during my 3+ years working with the company.
The Good: Articulating an Idea and Raising Money
From the time Jeff Stewart and I decided to press forward with Monitor110, we were very thorough and systematic in our approach to raising money. First thing I did before formally deciding to invest was to go around to potential customers - principally hedge funds and Wall Street firms - and get their reaction. We were able to show them the v1.0 system (the one we trashed, remember?), screen shots of the next generation application and our vision for the future. The general response: if you guys can deliver what you say you can deliver, I’d be very interested in the product. Good answers; just the answers a potential investor (me in this case) would like to hear. Little did I know that what we said we could deliver we actually couldn’t, and that it would take almost three years for the key constituencies (management, development, the Board) to figure this out. [The reason: buy-side investors wanted a system that would extract the nuggets of differentiated, unique information and present it to them in a form that was beyond clear, something so easy that they could look at it and literally make a trade. In short, readily actionable. Reality, unfortunately, did not line up with this expectation. The ultimate release was much more of a research application, not something that immediately generated actionable, monetizable data.]
Be that as it may, we didn’t know the execution problems at the time and the vision we were articulating was very seductive, to clients, to investors, to recruits, to ourselves. From a variety of client meetings we honed our pitch, created a presentation deck that we iterated on several times, and set about pitching a variety of top angels and VCs in Silicon Alley, Silicon Valley and Route 128. We raised $1.25 million from a group of angels, most of whom we considered strategic due to their either working at hedge funds, bulge-bracket Wall Street firms or were independent traders. This gave us the runway to make key hires, begin building the next generation product and to systematically tap the venture capital community. The pitch deck was no more than 15 pages, had lots of pictures, a handful of focused, powerful words and concepts, and a clear explanation of how we intended to use the money. While I’m no Guy Kawasaki, I know a good pitch deck when I see it and can sell it. Especially since I believed so passionately in the company’s mission and prospects. So my message to those seeking to raise funds is:
- Believe deeply in the mission and vision of the company; otherwise, no one else will.
- Use few words, many pictures and be brutally clear. If the audience doesn’t get it within 60 seconds, it’s tough sledding.
- Think of lots and lots of use cases and be ready to share them at will. This isn’t just for pitching; you’ll need this to understand the market opportunity as well.
- Pitch early and often. We learned so much from speaking to dozens of smart, insightful people. I think we would have failed faster and better and/or increased our chances of success if we had listened more.
- Hone the pitch on lower-likelihood prospects early and ramp up to the real targets after polishing the presentation and the delivery. The first bunch of times you will suck. After sucking for 5-10 times you’ll tend to get much, much better. There is no way to short-circuit the process; there is simply no substitute for experience.
The Bad and The Really Bad: Recruiting, Metrics, Lack of Focus and Cash Management
- Great team, wrong team
- Inadequate metrics
- Resources spread too thin
- Poor cash burn management
The Bad #1: Great Team, Wrong Team
I’d argue that Jeff and I were very good at recruiting. We hired great people. Smart. Passionate. Opinionated. Caring. Problem is, I think in retrospect that we hired many of the wrong people, not because they weren’t good but because they didn’t have the depth of experience necessary to solve the problems we need to solve. We took the approach of hiring “best athletes,” on the theory that super smart people can figure hard stuff out. For example, a few years back one of the teams that competed and placed in the DARPA challenge wasn’t from MIT, Stanford or CMU, but a team hacked together by a tech guy from an insurance company and a bunch of his buddies who said “We can do this.” They drew parallels between the DARPA challenge requirements and the dynamics of video games, and devised a completely out-of-the-box approach to solving the problem. And they proved that it didn’t take a bunch of rocket scientists to compete in a crazy robotics competition. This was our fantasy. We’d be these guys.
Unfortunately, our reality is that we were cracking monumental problems at the frontiers of natural language and statistical text processing, data harvesting (ripping and cleaning), entity extraction and real-time matching of terabytes of data. Instead of building a team that had experience in each of these areas and could attack the scale of our problems from the get-go, we staffed the company with bright people who had to learn the material first before attacking the problem. Because of this, it took much longer to understand the magnitude of our issues, time that could have been spent either solving the problems or taking a different approach. By the time we figured out the depth of our issues, we had burned lots of resources which left us precious little runway to get a salable product to market. Best of intentions. Worst of decisions.
The Bad #2: Inadequate Metrics
It took us a long time to build the next generation product. And during the development process, we didn’t do a good enough job creating metrics to measure our progress and the efficacy of the components of the system. And by the time we got around to measuring stuff (hard release cycles, data precision and recall, speed of throughput, processes around source expansion and quality, etc.) we were very, very late in the game. In retrospect, we should have placed a much greater emphasis on the creation and use of key metrics, and building greater accountability into the culture. It would have forced us to face into our issues early on, potentially enabling us to change direction before wasting scads of precious capital. Our lack of a metrics-driven culture let the science project live on - and on and on. We discussed the importance of measurement dozens of times but got derailed by the many technical problems we encountered. We completely lost the forest for the trees. This was no excuse. We screwed up on this front, big time.
The Really Bad #1: Resources Spread Too Thin
We initially had a vision of a single product: a dashboard. The entire company was executing against this vision. Problem was, the market was telling us that a dashboard was not necessarily what it wanted. It wanted a research product using the data underlying the dashboard. It also wanted the ability to access our data via a feed or API. Therefore, our laser focus on a single offering split into three. While we were trying to listen to the market, we only ensured that we would do nothing particularly well. And as mentioned in my earlier post mortem, the Board was never really supportive of the research business and didn’t really understand the opportunity posed by the feed business. What would have happened if we had only focused on one of the products? We either would have increased our chances of success or failed faster, both of which are superior outcomes to what eventually went down - a slow, painful death.
The Really Bad #2: Poor Cash Burn Management
Monitor110 wasn’t some biotech company conducting research and using money to get through clinical trials. It was a technology company trying to sell a service to a very defined universe. Our cash burn got way, way ahead of where it should have been given our revenues. Because, as we thought, the ultimate salable release was right around the corner. But it wasn’t. I can’t count how many corners that product was around, but it was more than five and less than ten. Yes, we were working to solve a complex problem. But before we went “all in,” we should have had much stronger signals from the market that it was prepared to buy what we were selling and at approximately the price point at which we intended to sell it. But because we were so concerned with disappointing our customers in light of the unexpected PR we had received, we really hadn’t gotten much market input since early in the development cycle. We were so convinced that the demand would be there once we got the product out that we just kept building - and building, and building. And when we finealy stopped building and said “Here it is,” they said, “That’s nice. Kind of. And that price? Too, too high.” Not happy words to ears that were poised to hear something materially different. And because of the burn level, it made staying in the game long enough to get the penetration needed to succeed virtually impossible. Moral of the story: test the market. Think. Design. Execute. Go back to the market. Repeat. Stop when the market says “I’ll take it.” And then start the cycle again.
Sure, there’s more. But this will have to do for now.
——-
——-
COMMENT:
AUTHOR: Tom O’Brien
EMAIL: tobrien@motivequest.com
URL: http://humanvoice.wordpress.com
DATE: 07/29/2008 12:28:23 AM
Thanks for the additional comments Roger - we are in a very similar business - and are always arguing about data vs. research insight - and about all the new products we COULD do.
This is really helpful because it is an additional (and interesting) filter for our big - and day to day - decisions.
TO’B
——-
COMMENT:
AUTHOR: Kristian Hansen
EMAIL: kristian@lindzon.com
URL: http://kristian.tumblr.com
DATE: 07/29/2008 02:10:53 AM
Roger,
More great commentary that everyone running or thinking about starting a business should heed.
Thanks,
Kristian
——-
COMMENT:
AUTHOR: Jed Christiansen
EMAIL: jed.christiansen@gmail.com
URL: http://blog.jedchristiansen.com
DATE: 07/29/2008 04:55:44 AM
Roger,
2 for 2! This is turning into a really useful and instructive series of posts on what can (and did) go wrong for early-stage startups. Thank you for being so honest and open about what so many other people might just want to forget. (Then again, I wouldn’t expect anything less from a fellow University of Michigan grad…)
I really hope to internalize these lessons for my upcoming venture.
Jed
——-
COMMENT:
AUTHOR: jack
EMAIL: no@spam.please
URL:
DATE: 07/29/2008 05:26:17 AM
Very interesting, thank you for sharing these thoughts. I had been following monitor110 as I find the idea very interesting, but am sad to see it failed.
My experience is that any IT-related idea is always much much harder to implement than it is to formulate. Estimating the true difficulty is only possible with hands-on technical experience in the problem-specific domain itself. Example: Web search. Naively, it’s just about 1) downloading websites, 2) extracting keywords from the websites into a database and 3) retrieving/ranking the websites based on the keywords in a user’s query. But to actually make it work, you need a Google filled with specialists in AI, linguistics, distributed systems, hardware design, data centers and much more.
Nevertheless, I hope you can extract/retain a lot of the knowledge you gained while building monitor110. This is often a loss when a project fails: the entire intellectual property is not wrapped up well and thus forgotten forever. Posts like yours try to make the key business learnings explicit. Technical knowledge can also be made explicit: If the software is designed well, core functionality can be spun off as a code library and can be licensed to others or open-sourced.
I wish you all the best in all your future endeavors. It is important that people keep trying new ideas - I have deep respect for anyone who tries to push humanity further. Here’s to the Crazy Ones. See http://www.youtube.com/watch?v=Dvn_Ied9t4M
——-
COMMENT:
AUTHOR: MIchael Lazerow
EMAIL: michael@lazerow.com
URL: http://www.buddymedia.com
DATE: 07/29/2008 08:23:40 AM
Roger-
Great post. You should consider turning this post mortem series into a blog of its own, or at least a major strain of this blog (or even a book for that matter). You are now working with enough companies that you can pull together and highlight lessons of both failure and success. You’ve definitely had your fair share of both — more of the success side, I would add! One of the key topics to be covered — hiring. Do you hire great people who can adapt, solve problems and kick ass? Or do you hire specialists? The answer is somewhere in the middle — you’re not going to hire a great person who doesn’t know how to build search algorithms if you’re building a search engine. But given how fast tides change at startups, having nimble doers is extremely helpful.
Anyway, thanks for the post. See you soon.
MML
——-
COMMENT:
AUTHOR: Matt Harris
EMAIL: matt@villageventures.com
URL:
DATE: 07/29/2008 12:57:44 PM
Roger,
Very good stuff in these two posts, thanks for owning this outcome and mining it for useful information. Too few of us investors do that. Kudos.
Best,
Matt
——-
COMMENT:
AUTHOR: Carles
EMAIL: gwbv@yahoo.com
URL:
DATE: 07/31/2008 06:49:53 AM
Hi Roger,
I’m sorry it didn’t work out the way you expected, even the hard work put into it.
Best regards from your cousin-cousin in Barcelona.
Carles
——-
————
2 years ago
| view comments
| Investing
