Michael Edwards

There is a growing movement around what people in the market call Consumer Directed Healthcare (CDH).

Without getting too wordy, this movement is getting amplified by the information asymmetry flip that Google has done on many industries, including health care. For instance, you can model a physician visit as an information exchange transaction. It used to be the physician had much more information than the patient. Google has opened up all of this information, so that a patient can spend 100 hours researching their condition and then come back with specific questions or concerns that the average doctor may not have been aware of. The patient is essentially moving the level of discourse from ‘industry average’ to ’state of the art’ as they have at their fingertips research articles published by leading medical researchers…

You can see just from this brief description how powerful health information systems health can be, yet at the same time, how far it can still be taken as doing ad-hoc Google searches is not the most productive way to do things, especially when the resulting knowledge is stored in the patient’s head and not some kind of health-aware knowledge management knowledge management system that future patients can use as a starting point.

John C. Goodman has written a good introductory analysis on CDH that you can find at the Social Science Research Network.

Dare Obasanjo has an excellent post up titled Sharing social activity streams across the Web: How Gnip fits in. In it he supports the move from pull to push notification, but questions the reliance on one provider, such as Gnip, as a single point of failure (SPoF).

It’s clear that there is a need for both pull and push notification. Right now the web 2.0 space is quite lacking in the push notification arena.

However, I don’t see much problem of a lock-in or SPoF issue because the Gnip behavior will be abstracted itself into proxy libraries that let you switch between data providers. You’ll be able to run your own Gnip type thing if you want to move away, or switch to any alternate provider. You’ll see a federation of these things, so, for instance, Flickr will co-locate a Gnip type server in their data center that will provide similar services. As an application developer you will configure your remote data transport system in a kind of ring-down way where you try Gnip first, Gnip clones, then per-service Gnip-like servers ala Flickr’s colocated service, then per-service direct API’s in case of larger faults. As the same time, all of this will be happening in a ‘virtualized’ Gnip service running on your own servers.

What you’ll pay for is bandwidth, QoS, etc. It will be a commodity market where the top players provide extended services that larger startups/companies demand such as reporting, global fault tolerance, application optimization services, etc.

The opportunity here, especially for Gnip, is to be the first mover that establishes not just the standard, but signs up big companies in a strategic lockup that reinforces their position and standard. Everything is still open, but it will be harder for new entrants to propose alternate standards and methods. A well capitalized Gnip will be in a position to have enterprise IT departments make long term (5-10 year) technology bets that include lucrative support contracts. Scrappy me-too competitors will have a harder time making this happen. 

Along these lines, I suspect that the real dollars are going to be found in the thousands of Salesforce API implementations. Twitter and Flickr are sexy, and might get pulled into Salesforce, but in general there are way more many internally developed applications that need this kind of thing than there are web 2.0 startups willing to pay enough to make it a business. Ditto the above on broad B2B data transports — the type of stuff that has replaced EDI, including the EDI over XML derivatives, could really utilize a service like this. The best way to engage all the startups is actually to open up the Gnip system more to allow 3rd parties to flush out the support of APIs to include everything on the web.

Under the scenario above, it makes sense to look at companies like Cisco, IBM, and Salesforce as potential suitors.

Cisco’s take would view this as part of the network stack (ie, Gnip is a web 2.0 API switch), and additionally would see its potential to increase bandwidth usage on corporate networks by making it very easy for people to consume more data than they do now. This is important — while Gnip might reduce the total data bits that would have been transfered by a factor of 10-100 to 1 or even more, it ultimately has the potential to increase the absolute aggregate number of bits people decide to actually transfer due to ease of use, etc. More bandwidth consumed means more Cisco routers sold.

For IBM, it is a consulting play. Every cash register is going to connect to multiple services in the cloud and someone needs to be there to do the integration and hold the client’s hand. If you can charge other companies a ‘gateway fee’ to get visibility on the cash register then that is an additional revenue stream. There’s nothing stopping IBM from charging a store to access Twitter while also charging Twitter for access to the store because in this scenario IBM owns the platform which is the cash register; not the store, not Twitter, and not an API aggregation service.

Re: Salesforce, it’s the obvious industry player to do this given their incredible success of business use of their API which they estimate will far exceed use of their user facing web application. Helping people connect more services to their Salesforce accounts increases utilization, dependence, and thus lock-in for Salesforce. Additionally, it would allow Salesforce to engage the consumer centric web 2.0 community, potentially developing a new source of leads for their business services. The danger is that Salesforce may choose to release their own Gnip like service.

I guess I have to mention that it would make sense for Amazon to do this also as it is a perfect complement to their web 2.0 platform play they are executing against.

Internet startups can learn a lot from the open source community. Their mantra of "release early, release often" is not heeded anywhere close to the amount it should be. Some of the most successful open source projects, such as the Linux kernel release dozens of versions a year. It’s interesting because they almost never hold themselves to checklist items the way the agile development community does. Rather, they work hard and then just release stuff. Some of it works, some of it doesn’t work. No particular planning on the level of granularity that some people do.

However, their speed allows them to iterate more. More iterations equal more evolutionary cycles. More evolutionary cycles allow them to figure out what works faster and gives users the confidence that the project is active. Active projects hold the promise of fixing the bugs or issues you have ’soon.’ So, people are willing to deal with less than 100% quality if they believe that things will get better.

The same is true of most companies with top notch customer service. People become loyal to them because they believe their problems will get fixed. It has nothing to do with a lack of problems. Buyers know there will be problems and choose a vendor that will help fix them.

The equivalent in the startup world is a company that listens to its customers and responds very quickly with fixes and feature additions that unfold as a cohesive story that its customer community begins to identify with.

Twiddla is a very slick social whiteboard and wireframing system. Totally recommend it. Very easy to share whiteboards with people.

http://www.twiddla.com/

Medical procedure checklists are considered to be one of the most important clinical advances this century.

While everyone else is complaining about the inevitable doom sure to afflict startups across the board with death I’ve been rejoicing.

Imagine you have a big office building that’s hard to walk through because there are a bunch of people in the hallways.

Now imagine if those people leave the building.

Office is still there. And it’s there to serve just you.

What’s the problem with being the only one on the interstate highway? Wouldn’t you rather that?

The bigger the pipe, the less stuff in it, the more room you have for, well, you and whatever you want to be doing.

You can flip this into a gigantic land grab opportunity. People are walking away from opportunity. That’s fine. The opportunity is still there. We’re living in amazing times with some incredible megatrends propelling us further. Think about Moore’s Law of chip speed doubling every 18 months.

For some reason Moore’s Law hasn’t cared about anything in the news in decades.

There are hundreds of Moore’s Law analogs in different industries. They are interconnected in gigantic evergrowing feedback loops.

So, one of them is wilting a bit…

Big deal.

The biggest boom starts from nothing. I’m seeing DOW 20,000 in the future, so why wait for it?

Product diagramming is important in product design as well as explaining to the world what your product does on a technical level which is especially good for technical marketing efforts. Additionally, it’s sometimes useful to diagram competitor products as part of your competitive analysis routines.

I’ve started collecting examples of different product diagrams on the web. If you know of more please let me know. This list is just the beginning…

• Google Spreadsheet Features
• UltraEdit
• Industrial Grease Regulator

The latest Vator Box on Vator.tv features Esther Dyson analyzing Careflash and Yapta. Check out her analysis along with the Liquid Scenarios Minute to understand the exit scenarios possible with these two companies.

The interesting thing to me is how Careflash can grow to a significant size while only targeting its home market of Houston, Texas.

Below is a list of web based tools I have or am using for competitive analysis. If you know of a product not on this list that should be send me an email and I’ll update it.

• RivalMap
• StartPR
• Trackur
• Compete
• Quantcast
• Quarkbase
• Google Docs
• Office Live
• filtrbox
• Google Alerts
• Radian6

Update (2008-10-09): Added Radian6

I’ve used a lot of tools to do competitive analysis over the years. One of the big problems is that much of it went over email. It’s very easy to loose track of email conversations, very hard to organize all the content on a given subject. Furthermore, there is not a natural way to convert it into higher fidelity docs, charts, and presentations.

Another more general problem I’ve run into is putting lots of time into textual analysis. The problem here is you don’t have data that lines up with other data. You can’t group, sort, chart it, etc.

About a year ago a tried going the custom database route. It started off as a Google Doc, then Google Doc Spreadsheet. Finally, it ended up as an Access database. Access was nice in that it would easily spit out reports, but I ended up getting bogged down in designing screens and dealing with the best way to display and segregate certain data structures.

At the end of the day I could print out a report that was 100 pages long, let it land on a table and people we generally impressed at the amount of attention to the issue.

However, only one person could work on it at a time, it was relatively a static custom system, and much of what it was tracking was textual data, either in the form of hard facts or opinions. We stored this and could show it, but couldn’t analyze it using computational methods, couldn’t transform it into charts, etc.

From there I tried importing the data into some of the new web based database applications like Dabble. Dabble is an awesome product, but I again dealt with the inability to work with some of the data using the UI controls I wanted. Additionally, there were often times of slow page loads and I think the UI navigation in general is somewhat confusing especially for less tech savvy people who may have valuable insights to contribute.

All of these tools have their uses, but in the next set of articles on this subject I’m going to be reviewing a new generation of collaborative tools as well as a highly focused quantitative computational modeling approach to the problem — one that you can use to produce a dynamic ‘living’ model of your competitors, market position, and even overall business dynamics if you so choose.

Christopher Allen at Life with Alacrity has written an insightful post on group size and dynamics. He specifically looks at size thresholds that act as breakpoints beyond which groups generally have to behave very differently than when they were smaller.

This view of group size dynamics has a lot of potential implications for all kinds of things including management science and more specifically/interesting to me: startup dynamics. Namely, Christopher’s article can be treated as an outline that can be mapped to company size and funding levels, amounts of management required/deployable, liquidity scenarios, etc.

Choice quotes:

…If a community is too small you’ll often have insufficient critical mass to sustain it. Conversely, if it’s too large you can end up with a community that’s too noisy, too cliquey, or otherwise problematic. These optimal and sub-optimal community sizes appear in strata, like discrete layers of rock. For a community to advance from one strata to the next often takes immense energy…

…Clearly, as we step up toward higher group thresholds, more and more time is required to simply keep the group going. You see this in depictions of mafia life — in the TV series The Sopranos a lot of time is spent dining, hanging out, and drinking together. That is part of that 42% social grooming time required for that intense of a survival group…

Liquid Scenarios created a model of the bailout bill that shows how its first year interest costs could have been financed by appropriate warrant coverage and rules. It’s amazing that an application that venture capitalists use to manage and optimize their portfolios can scale its capitalization and security holder modeling engine to 100-1000 times its usual size to model an entire country’s financial system.

Here’s a link to the official press release:

http://www.prweb.com/releases/warrant/liquid/prweb1405774.htm

I just stumbled across a great set of articles that Mark Davis of DFJ Gotham Ventures has written on the fund raising process, including good coverage on managing capitalization tables and valuations:

• Sharing Your Cap Table
• Option Pools
• Keep Your Cap Table Simple
• How A VC Can Change The Cap Table
• VCs Want Key Team Members To Have Significant Equity
• Why Liquidity Preference Exists

Liquid Scenarios has created some incredibly cool search technology that creates predictive acquisition models and competitive portfolio dashboards in seconds using data found on the web.

If you’re a startup founder this is important because for the first time you can quickly understand the significant exit breakpoints of competitors and develop strategies to provide better value to potential acquirers.

Likewise, in seconds you can understand the median purchase price of comparable companies and run breakpoint analysis on your own company comparing the relative purchase price of competitors to their traffic levels (if you’re an Internet startup) to understand your current value.

The quicker you can iterate and evolve over these models the easier it is to understand not just your current competitive market dynamics, but those of adjacent markets. Maybe you have a great product and reasonable traction in a market with historically low exit values or volume. In seconds you can do what-if’s on nearby markets…

How much is it worth to know that if you re-branded your existing product and came out with a follow-on tuned for a different market that you would be set for a 10x greater return than staying where you are?

This summer I attended the AlwaysOn Stanford Summit with Liquid Scenarios. Below is a video I appeared in that was presented to an audience of a few hundred plus several thousand webcast viewers.

These guys turned protein folding into a game for people to play. When you fold proteins you’re results go into a database that is used to create real cures to real diseases. It turns out that one human still has more computational power to perform this kind of problem than all the computers in the world . . . So if we get thousands of people to do this great stuff will happen.

You could probably teach 1st graders more about protein structures than most college students know if you just had them play this game for a while.

My blogging activities go back to 2004, yet I haven’t consistently posted here through the years. That’s going to change.

I’ll be talking about the convergence of evolutionary medicine, green tech, startups, finance, globalism and Boulder, Colorado.