The Rise and Fall of Swivel.com

Swivel is about data

Earlier this summer, the visualization website Swivel.com disappeared from the internet. To find out what happened, I tracked down and interviewed Swivel’s two founders, Brian Mulloy and Dmitry Dimov.

The following is an edited combination of two interviews I conducted separately with Brian Mulloy (who served as CEO for two years) and Dmitry Dimov (Chief of Product), interspersed with some additional thoughts. My main goal was to learn more about the thinking behind launching Swivel, what went wrong, and what we can learn about commercial visualization (and “visualization-as-a-service”) from what happened to the company.

YouTube for Data

Kosara: How did you get started?

Mulloy: Dmitry and I started in enterprise software. We were selling multi-million dollar software licenses to financial and telecommunications companies to help them integrate their systems. We wanted to solve the integration problem through software as a service on the Internet, rather than running it in their data centers. We also wanted to visualize the data so that people could understand that the integration had value. Seeing all the data for your business in one spot would be really valuable.

As we were getting close to launching the product in 2006, our advisors, Halsey Minor [founder of CNET] and Ron Palmeri, saw the YouTube acquisition happen for $1.6bn. They started to wonder if there was a way to make our product into more of a general-purpose visualization story. In the end, we launched Swivel as what bloggers called “YouTube for Data.” As you can
see, it was sort of a random walk into this space.

Kosara: Was the moniker “YouTube for Data” actually your own term, or was it coined by somebody else?

Mulloy: No, that was Michael Arrington’s [of Tech Crunch] term. He saw a demo and that’s how he thought of it, because we had the commenting and what, at the time, were the canonical web 2.0 features. We never actually used it ourselves.

Public Data vs. Private Data

It’s interesting to see this switch from very a narrow focus to a broad visualization platform. They took it even further by inviting organizations to feed them their data directly. These datasets got the Official Source seal, which gave users the confidence that the data had not been tampered with.

Kosara: Why were you interested in public data, like what you had under the Official Source program, and the vision you presented at the OECD World Forum 2007?

Mulloy: We wanted to get people engaged with data so we could build a better social visualization product. We hired Sara Wood, who was responsible for web technology at the World Health Organization (WHO) in Geneva. She had blogged that Swivel was cool, and she had an idea to do something similar. I reached out to her because we wanted someone to help us with big data sets and to figure out how to make them more usable. Later, the OECD actually contacted us to put their statistics into our product and we were blown away by that. That was the genesis of the Official Source program.

Dimov: Private data sources are sensitive, so we decided to focus on data that was public, but still relevant to our users. There was no problem with private data anymore, but it was still data that people cared about. I may not care about somebody else’s public data, beyond just general interest, but I think a lot of people want to look at data that relates to them.

Swivel as a Business

Kosara: How large was Swivel? How many people, how many servers, how much money?

Mulloy: It was really cheap. We were five people at launch, and we only ever grew to twelve, and some of those were interns. We only had a single rack of servers, ten machines. We weren’t streaming terabytes of video, the datasets were quite small, so the infrastructure requirements were not huge.

We got a Round A funding of two million dollars. By the time I left, we had spent about three to four million dollars on the idea.

What Went Wrong

I was surprised by Swivel’s disappearance not just because I had assumed that they were doing well, but also because if I’d had to pick a winner between Many Eyes and Swivel back in 2006, my money would have been on Swivel. When I wrote my original review of the two sites, I noted Swivel’s plan for paid accounts and Many Eyes’s lack of a business plan.

Kosara: You seemed to have a very clear idea about how to turn a profit by having paying users in addition to free accounts and the general public. How many paying users did you have in the end?

Mulloy: It was single digits. To be honest, many of the things you’ll see with Swivel are more of a cautionary tale for a first-time CEO like me than for the space as an opportunity.

Somebody’s going to solve this problem, and they’re going to be a really, really big company, it’s just a matter of time. We had a notion that we could do the revenue model in a way that would be purely web-oriented and self-serviced, at a really low cost of sale; following the model of 37signals. I questioned this a few times, but my investor didn’t want a huge enterprise sales force. You probably can build a profitable business model that way, but looking back now I think we would have served the company better by hiring a few sales people, really understanding what kinds of problems people were having with their data and going after it as a real money-maker.

Kosara: So what went wrong?

Mulloy: It all points back to this: the CEO was incompetent, or not as on the ball as he could have been. I couldn’t have said this easily a year and a half ago, but now with more time between me and what happened, I think it was just a number of mistakes. We had phenomenal people, phenomenal investors, really talented people. When you look at what the guys and gals have done since Swivel, they’re all really interesting, fun projects, that are making a difference. So I think I didn’t orchestrate the business as well as I could have, and made a lot of little mistakes that added up to not being the success that it probably could have been.

I often talk to people, and they ask me: what did you do at Swivel, what can we learn from your mistakes? I’m really still very optimistic about the opportunity, I think a really good executive team with a lot of focus will make it into a really big business. It’s just, it wasn’t me.

Dimov: It had the feel of a big idea. There was a lot of interest, and there are so many things you can do with data, there are really no limits. But we weren’t able to describe what the actual uses were: this is what you can do, this is why it’s useful, so people would use it. We also weren’t able to prioritize our features, and we had very limited resources in terms of how many developers we had, etc.

We had so much good feedback from people who touched Swivel, I don’t think it was an issue of not being useful to people. There was such a huge variety of things to do, that as fresh entrepreneurs, we had difficulty saying: here’s the one thing we’re going to do. And many companies, even when they focus on one thing, realize that getting different kinds of data in from users is just really difficult.

Single digits. I double-checked this with Mulloy, they really had fewer than ten paying customers. I find that shocking, simply because of the statistics: Swivel had (tens of?) thousands of registered users, but fewer than one in a thousand bothered to pay for the service? Unbelievable.

Is There A Need?

I asked Mulloy what he thought about Stuart Roseman’s assessment of what led him to shut down Verifiable.com. In particular, I was interested in the questions of what need people have for such a service, and how much effort they were willing to put into using it.

Mulloy: Our hypothesis was that there are a whole bunch of people who are not analysts or statisticians, or visualization experts, who would really benefit from seeing, and engaging with, statistics. And, if we made it engaging, they would. That was our starting point. So we tried features which we got criticized for, rightly or wrongly, like blinging our graphs.

I think what we learned, like Roseman is saying, that the interface is not that important, that there are analysts who are really good at tools like R, SAS, etc. and prefer to continue to work in those tools to do powerful things with datasets. And people who are not inherently biased towards working with datasets, they are not going to do it. Except for what they see in the newspaper, like USA Today or the New York Times. That may be the end of it.

Brent Fitzgerald, a former employee of Swivel (and one of the speakers on an InfoVis panel on Social Data Visualization that I organized at InfoVis 2007), nicely summed up the problem in an email:

I guess one of my Swivel takeaways is that visualizations are tools or lenses, and people will create visualizations out of need, in the contexts in which they are needed. It’s easier to start with the problem, then solve it, than the other way around.

Visualization Expertise and the Community

What struck me as odd from the very beginning was that none of the people involved in Swivel or on its advisory board seemed to have any background in visualization, or were known in the field.

Mulloy: We did have a visualization advisor, and we spent quite a bit of time with him; but I think he was so embarrassed that he asked not to be named on the website. He’s a very prestigious, award-winning visualization expert, but we couldn’t tell anybody because he was embarrassed.

Kosara: Do you think that “chart bling” (adding a background image to a chart) was the right kind of feature to focus on?

Mulloy: No, no I don’t. We got slapped around quite a bit for that, and in retrospect it was a mistake. We never claimed to be experts in the field of visualization, and we probably deserved that beat-down.

Kosara: Would more involvement from the visualization community have made a difference?

Mulloy: Yes, definitely. And after that [bling backlash], we reached out to some academics here in the bay area who helped us. The onus was clearly on us, and we would have been better off to get that input sooner.

Many Eyes and the Commercial Visualization Space

Kosara: Was the competition with Many Eyes perhaps a problem for Swivel’s business?

Mulloy: We were very impressed with Many Eyes when they launched, and I actually went to Cambridge to hang out with Martin Wattenberg and Fernanda Viégas. Quite frankly, they had a level of horsepower that we just didn’t have available to us. They’re truly brilliant thinkers, and they’re still doing things in this space. I can’t say enough positive things about what they’ve accomplished and what we learned from how well they executed everything, from the design to the data integrity, it was just all so well done. We learned a lot, and it helped keep us honest about what we were trying to do.

They weren’t just competition, they bested us right out of the gate. Which was a real bummer, because Dmitry and I spent the last five years before Swivel competing with IBM in our other job, and just when we thought we would be doing this web startup and never have to see IBM again, boom, one month later Many Eyes came along.

Kosara: Were there just too many visualization websites out there? Is there only room for one?

Dimov: I don’t think that there are too many, because there really isn’t one that’s useful to me right now. Swivel was useful, but then I was an insider, so I knew how to make it work. But I don’t think that anybody has built the right thing yet.

The Final Months

Both Mulloy and Dimov left Swivel a while ago, so they could not comment on when and why the final decision to pull the plug was made. Dimov left in January 2008, only about two years after the site was launched; he is a co-founder of Armageddon Energy. Mulloy handed over the CEO position to Visnu Pitiyanuvath about two years ago, and originally stayed on as an advisor, but then also left; he is now with Apigee.

Kosara: Why did you leave?

Dimov: We were stuck in neutral for a while. We launched in 2006, we had a lot of excitement in 2007, had the OECD approach as we went to Turkey [OECD World Forum in Istanbul, 2007], Google was interested, and there was no shortage of people wanting to invest. But we still were just a few developers, and we were really going around in circles; we didn’t have a clear idea what version 1.0 of Swivel was going to look like.

This is the reality between a prototype and a working product: there are so many details that go into turning an original idea into a product, most people don’t realize that. You can spend a lot of time on that in fact, to the point where a year goes by and all you’ve done is that you’ve fixed a bunch of bugs. So you can either scale back the features and only support a limited set of functionality, or scale up your team to support all of those things. And just for the data management you can spend a lot of time.

We had a great team, developers that could do anything, but it all takes time. I’m certainly wiser now, to know how easy it is to build a nice façade, and how much more work it is to really develop these features and make them truly functional.

Kosara: Right now, Swivel.com is completely dead, there’s not even a placeholder page. Do you know when that happened and why?

Dimov: That was a surprise to me. I’m busy with my startup, so I rarely check in. I know that a few months ago, there were two developers who were essentially just keeping the site running. As far as I know, back in June or July, there was nobody actually there anymore, it was just a few servers that ran the site and everything was working as if the company was still there.

Mulloy: That was actually news to me too. When you reached out to me, I went to the site, and saw that it was gone. I really don’t know what happened. To keep the data center running, somebody has to pay the data center vendor, and I guess they just stopped paying for it. It’s strange.

(I am trying to contact people who were there after Mulloy and Dimov had left, in particular the VCs and the CEO, Visnu Pitiyanuvath.)

The Future

Kosara: Will there be a YouTube for data that’s the same kind of success as the current YouTube?

Mulloy: No. I think it’s a question of context. 99% of what made Hans Rosling’s health data so interesting was Professor Rosling himself. He would swallow a sword that was on fire as part of his presentation. I think that that context needs to be there for people to engage with the data. On YouTube, you don’t need context, you just play the video and then you watch the cat jump around and that’s all you need.

Kosara: Do you think that efforts like Apps for America and other public-data initiatives will make a difference and continue to attract developers and users?

Mulloy: Speaking from my heart, I would hope that these things continue, because this is something that really needs to stick if we want people to engage with data, and if they want their government and their world to be any better than it is. But speaking from my experience and with my brain, I’m afraid it’s going to be a fun project on the supply side, but there isn’t going to be any demand there. And it doesn’t matter how great the mousetrap is that they build, I don’t think it’s going to last.

Kosara: That seems really gloomy. Back in 2006 and 2007, there seemed to be this gold rush around visualization for the masses and public data. Do you think we’ll see anything like it again?

Mulloy: It’ll come back. It was a flowering. kind of a golden era. But it’ll come back. I’m optimistic that there will be another flowering like that.

Kosara: Is there a need for visualization-as-a-service? Do companies want to hand their data over to a third party to analyze it?

Mulloy: No. I don’t think they do. But there is one thing that changes this, and that is mobile devices. We’re going back to a client-server model, where mobile applications need to access rich data to be of value, and they do that through APIs. Companies will need to open APIs if they want developers to build interesting apps based on their data. And once these APIs are there for mobile phones, you can access that data from anything, like a television or a game console. And we’ll be able to build visualizations with that.


Hat-tip to John Bolton (of the Arizona Daily Star), who originally alerted me to the fact that Swivel was gone. He was an avid user of the site (like many journalists, it appears), using the embedding features that are now also gone, of course. He nicely summarized the problem for users:

These free tools on the web are easy to use, easy to lose. A cautionary tale.

25 responses to “The Rise and Fall of Swivel.com”

  1. Jan Willem Tulp Avatar

    This is a very interesting interview! I also read the “The End of Verifyable.com” post a while ago, and as I am starting my own company (called Lookalytics) right now, the biggest lesson I have learned from Swivel.com and Verifyable.com is that it probably isn’t the right time to create a business around a web-based general purpose tool, and solving a problem first should be the primary focus.

    Although I hope to make a business in Data Visualization, I am still doing some market research for my own company, to see where the problems are that could be solved using Data Visualization. And the stories of Swivel.com and Verifyable.com do help with that.

    My general impression is that to be successful as a Data Visualization company, you should do one of these:

    • do custom projects, like Stamen, Vizzuality, MonsterSwell, etc.
    • make a Business Intelligence tool so that people can analyse their own private data, but there is a lot of competition here, and lots of large established companies
    • build a product around some specific kind of data and use Data Visualization in the product, but don’t ‘sell’ it as a Data Visualization tool. For example, if you’re a project manager, and you want to measure performance of your team and projects, your more likely to have a ‘project management problem’ than a ‘Data Visualization problem’. So a project management tool will more likely fit the problem than a general purpose Data Visualization tool.

    So even though I really feel tempted to build my own general purpose Data Visualization tool myself, just because it’s really cool, I keep looking for some real problems to solve first I guess.

    Anyway, the end of Swivel.com, as well as the end of Verifyable.com, has taught me some very valuable lessons. So Robert, thank you as well for these informative posts!

  2. Carlos Scheidegger Avatar

    (This is a great story, thanks for tracking down Mulloy and Dimov and running these interviews. I posted the link on Hacker News; that’s a great place to read about startup stories like this)

    Like Jerome said, the space in which Swivel was living is getting pressured from below from the low-level Google charts tools, and from above from Tableau Public. Both companies can afford to not make money on these products. So unless you have a killer product that’s sufficiently different from these two for which people cannot line up to pay you fast enough, your company is dead on the water.

    Unfortunatel, I also ultimately agree with the pessimistic thoughts of Muloy. I keep going back to thinking that visualization is, ultimately, a means to an end. So if you sell visualization, it’s hard to move away from an indirect sales pitch (“you buy our tool/service, then you will be able to visualize your data, then understand $foo better, then make more money”). In this situation it might be a better idea, business-wise, to pitch the ultimate point directly (“we’ll teach you all about $foo”). But then, you’re focusing on a particular type of data analysis, and you’re no longer a visualization company.

  3. Xiaoyu Wang Avatar

    This is a very insightful interview! Thanks Robert! I think the take away fro this is “What is actually a correct biz model for Visualization?” I have be dwelling on this motion for two-years since the first thought about having another startup on my Ph.D. research. It sounds cool and really useful. But the biz model is just hard to nail down.

    The one thing you interviewees point it out is “[For visualization], there is problem first then a solution”. I agree with them however, I see this as a problem resulted by public vs. private data.

    Public data invokes interests, at best, while private data demands investigation. For one reason or another, private data in the enterprise or company world encapsulates information that people wants to dive deep in. It may shows a profitable pattern for a company, or it could indicate a potential revenue path. In this sense, private data relates itself more to profit and demands investigation.

    On the contrary, public data may lack the soul of revenue generating. It is interesting to look over public data, but people would eventually lack motivation.

    I think, like I am trying to approach, is a startup that mixture about private and public data for investigation uses. This can be and should be achievable based on the experience we had with Bank, with DARPA, DOT, DHS and etc.

    While I like the idea of making visualization solely for masses, it is more startup-ish when we dealing it with more specific private sectors and interests.

    Anyhow, again, great interview, I always wanted to discuss this with someone in the industry. :) Next time, bring me with ya, you know I am just a lab away. :D

  4. Anthony Calabrese Avatar

    Thank you for getting this information out there!

    There are some important lessons in here for entrepreneurs, journalists and dreamers of all stripes.

    First and foremost, a great idea is just the beginning. Getting wide recognition – just baby steps. You’ve got to seal the deal and show people how and why they will use your product – whether it is a service, a visualization tool or something else.

    If you are for profit, you’ve got to sell something that people want to use so badly that they will pay for it. I was shocked to learn that the paid users were in the SINGLE DIGITS. Was the service that expensive or are people just not that into what they had to offer?

    Swivel was one of the first visualization tools that I had ever used. I had my gripes with the user interface and the occasional bugs, but it was good enough to get the job done. They are still prominently featured on the Minor Ventures homepage: http://www.minorventures.com/

    I am sad to see them go without any warning and I know lots of Swivelers are fuming now that an untold number of visualizations are dead and have to be rebuilt from scratch.

    You quoted John Bolton and his words ring true now more than ever: ‘These free tools on the web are easy to use, easy to lose. A cautionary tale.’

  5. Jerome Cukier Avatar

    I am surprised that Brian is taking all of the blame. I met him in 2007 when he came to OECD and he really impressed everyone there with his drive, his vision, his focus.
    The problem was more in turning the Swivel idea into a business. Inherently a tool that generates charts out of data has costs but no revenues. The long-term goal was to let companies easily analyse their data and share these insights inhouse. This is still an unmet need, but while business leaders are aware they have too much data to analyse they don’t think of this situation as a data visualization problem, and so they are not looking for a data visualization solution.
    There were two other problems with Swivel. Building the social community around data was very resource-intensive. Swivel was launched pre-facebook, pre-twitter (well facebook existed but not like today) and they were not able to use existing communities, based on existing social networks. Like, they could have powered wikipedia (which btw has a serious chart problem) or other similar tools with great following.
    The last problem is competition. Today the closest equivalent to swivel (ie the easiest way to create and share a chart online out of data) is google chart api. this may not be a perfect tool but it’s supported by google. The other tool I’m thinking of is Tableau Public, which is very advanced and which is one part of a (profitable) business plan. So it would have been difficult to stand against these 2.
    I really wish Brian and Dmitry best of luck in their ventures and the best to the rest of Swivel staff too.

  6. Sara Wood Avatar
    Sara Wood

    With any startup, there are ups and downs, small mistakes and big mistakes. I joined Swivel as VP & Chief Data Officer early on and was completely blown away by the talent and intelligence of the team and the vision of the founders – Brian Mulloy in particular. I think everyone involved truly had (and some still have, as I do) the strong belief that data and tools for understanding them are the next place for the web to have true impact.

    I chalk a lot of the shortcomings of Swivel to timing, not lack of vision, interest by early-adopters, mass-market appeal or talent of the team. Web-based technologies needed to assemble complex interactions (but in a simple UI) for real-time data analysis are still maturing and people’s comfort level in sharing data on the web is still evolving as well (Particularly business or research data – this was before “in the cloud” was a phrase middle managers had ever heard of, and frankly are still afraid of. And last I checked, even tax-payer funded research data sets are pretty closely guarded).

    I left Swivel to run Product at Flickr and am now overseeing Product strategy at PLoS, a non-profit dedicated to making more scientific and medical research open access with open data supporting it. Through Swivel I met an amazing community of like-minded folks (like the Robert, who runs this site) that, on a personal level, make Swivel a huge success for me. People with whom I hope to have healthy discussions into the distant future about the intersection of data, the web and the impact on society.

  7. Martin Avatar
    Martin

    Thanks for the interesting interview. Working as Ph.D. on a knowledge-assisted visualization tool, the insights of this interview support my understanding in using formalized domain specific knowledge to solve domain specific problems.

  8. Joe Hellerstein Avatar

    A few thoughts as a technical advisor at Swivel:

    1) Brian Mulloy is a class act, as you can tell by the way he keeps “taking one for the team” in a lot of these quotes. The buck stops at the CEO, and good for him that he won’t pass it. I’d work with Brian again with pleasure; he’s a remarkable leader. The whole team at Swivel was impressive.

    2) It’s easy to look back and critique, and ordinarily I wouldn’t do so in a public venue. But I’d like to encourage more work in this area, so the following is intended in that constructive spirit.

    In addition to some of the discussion above, many of the takeaway lessons from Swivel for me were about engineering execution. I think this explains the handful of customers: by the time the feature set was sufficient to address business needs, it was pretty late in the game, and the excitement had subsided on a lot of fronts.

    The first public version of Swivel was good enough to generate lots of attention; it was a really smart first cut on a new problem, and attractive to dive into. A lot of its appeal came from some really nice technology developed by the initial team. They independently discovered and shipped a bunch of technical ideas that some readers here may know from the research literature, including a declarative visualization language, automatic graph generation from data, type induction from text, and so on. I thought they showed a lot of taste and wisdom in avoiding exotic visualizations, and instead focusing on things that could grow community enrichment of public data, like hands-off vis generation tools, and an effort to welcome both official sources of data and random driftwood from the Internet.

    Agile development and new tools like Rails worked very well for them in that early phase. But as the site started to mature towards usability for a “private” business model, the clever technical ideas became less important than some of the nuts and bolts, and sometimes got in the way. The dev cycle started to bog down in complex exercises in rewriting the code — in many cases more or less from scratch. I suspect that an experienced engineering manager would have seen these issues early and steered the site to a much more tactically useful state much earlier on. The engineering team was absolutely first-rate and I respect them all — even moreso now that they’ve been through this process. But I think some old hands in the room a few years ago would have made a big difference.

    It’s important to remember the context when reading this. Swivel was coming up during the Web 2.0 heyday, when many things were getting off the ground on the cheap using a young talent pool of sheer smarts and hard work. Unfortunately I think that approach was better suited to YouTube than “The YouTube of Data”. Data analysis and data vis software has technical and design issues that are much more interdependent and multidimensional than most early Web 2.0 offerings. It’s the kind of stuff where tradeoffs have to be made just to get to basic usability, so you want a mix of sheer engineering firepower with product and engineering management experienced in the art of compromise. With a bit of that, I think Swivel might have found its way.

    I remain optimistic about web-based collaborative data analysis and visualization … it’s feels inevitable to me.

  9. Ben Warner Avatar

    One of the potential audiences for Swivel was the network of practitioners in the community indicators movement — people and organizations dedicated to the democratization of public data and turning local data information into usable sets to create data-informed public dialogue and decision-making.

    I remember at the Community Indicators Consortium conference in 2007 the excited buzz about Swivel as a potential resource to enable the communication of local-area statistics with the general public. Enrico Giovannini from the OECD was particularly positive about the possibilities, and a number of community-based organizations took notice.

    Many of us watched the development cycle and waited, I suppose, for the point at which the ease of use and the visualization tools would meet local needs. In the meantime, a number of other data visualization tools — not just Many Eyes — were also competing for our attention. Most of us work on very limited public or not-for-profit budgets, and so had an (understandable) hesitation of being early adopters of a particular system.

    My organization turned to a different data visualization system for transmitting our information to our audiences. The solution we chose is not perfect, nor does it have all the features that Swivel offered. The primary concern for us was the sustainability of the tool(s) we used — it’s hard to train non-data-oriented audiences in how to be comfortable with using data, and we didn’t want to risk having to switch systems and retrain. That meant we had to use a system where we had physical control of both the data and the data display technology, so that if/when a company went away, our communications strategy with the community would remain unaffected.

    I still see a large audience that is desperate for good data visualization tools to share local public data with broad audiences for dialogue and public policy decisions. The challenge is that this audience probably isn’t going to invest the kind of resources necessary to create the ROI anticipated by these kinds of start-ups. I’m looking forward to someone creating a business model that takes into account both these kinds of audiences and their price points.

  10. Halsey Minor Avatar
    Halsey Minor

    Swivel will be back up. We have no intention of stopping development or losing all the good learning the great people at Swivel have helped to generate. It’s a hard problem we are going after which is why their is no successes in this space as all who have worked on Swivel know. Maybe with some perseverance we will get lucky and get it right, and maybe I will end up enjoying trying to solve the riddle.

    Regards

  11. Enrico Bertini Avatar

    Hi Robert, great interview!

    Anyway I want to make sure we are not taking the failure of swivel and verifiable as the failure of visualization! There are two aspects to take into our mind:

    (1) Failures are good! We need lots of failures to find the next big thing
    (2) Visualization is thriving everywhere and it takes time
    (3) Swivel and Verifiable are failures of a specific business model (or its execution?) not of visualization.

    Let me iterate over some recent successes of visualization: Tableau is a very well respected company and growing a lot. It is in the Gartner Magic Quadrant 2010 (http://www.tableausoftware.com/gartner-magic-quadrant-2010) together with Spotfire. Infosthetics and FlowingData have thousand of readers. Eagereyes, even though it requires a more reflective audience, has a good bunch of subscribers isn’t it? Visualization is alway in the news and the New York Times is an established mainstream player. And oh yes, VisWeek is getting a big and respected conference. We are making some noise!!!

    What I really think is that the whole model of fruition of visualization on the web is really challenging. Swivel and Verifiable were nice but I visited them a couple of times for curiosity and never went back. Same for ManyEyes. Maybe it’s just this model that doesn’t work. And why should we insist on that model? This is not clear to me. Also, a very large proportion of businesses on the web don’t even try to monetize through the service they propose directly. I am wondering if Swivel just tried to monetize in the wrong way. If you have thousands of followers not necessarily you have to let them pay for your service directly.

    I totally agree with Jan. The best model for selling visualization is to consult people for custom visualizations like stamen and the like do. I am not a big fan of any generic visualization tool anyway, despite the success of some companies. I still think that the big bulk of needs come form small custom project that solve the specific needs people have. We always have to remember this, the large majority of people don’t want to stare at the screen and contemplate how cool a visualization is. At least those ready to put big bucks on it. **People want solution to their problems.** This is a simple timeless advice that we all seem to forget from time to time.

  12. Alex Kerin Avatar

    To repeat what others have said before, but it bears repeating, I just don’t believe there was a business model that existed. You can’t mention Tableau in the same sentence as Swivel. Swivel took data pre-formatted and made some charts. There was some interactivity, and the links to grab data, but there was no business problem that was being solved.

    No one turns around and says “if only there was a way to get this data I’ve already summarized and worked on in a chart format on the web”. They just draw it in Excel, take a screenshot and whap it on the web. Or, more likely, they email Steve in marketing.

    On the other hand plenty of people say “I have a file with 100,000 rows of data, I wish I could understand it better, and then tell other people about it”….

  13. Si Chen Avatar

    This sounds like great technology, but the market for it really just does not exist yet. For example, a few years ago a trade journal (either Managing Automation or Manufacturing Business Technology) published a survey that showed that half the manufacturing companies in the United States don’t even have an ERP system. Most retailers have multiple CRM systems and databases with duplicate and inaccurate customer data. Just getting the data together will be a huge challenge for years if not decades, so data visualization unfortunately remains a bit “pie in the sky” — which is too bad, because it really is very cool stuff.

  14. jon Avatar
    jon

    This is an amazing interview. It’s very useful to other entrepreneurs to have very detailed post-mortems like this. The fact that Mulloy so willingly admits mistakes is really refreshing, and it shows he’s learned a lot from this and will surely win another one down the road.

  15. Halsey Minor Avatar

    I don’t want to debate whether Swivel will succeed or fail. My philosophy is you are more likely to succeed the longer you try. Other people may have different approaches. I do have 2 important facts that guide me regarding the future.

    In 2004 I built web publishing software called “Story Server.” I sold it in 2005 to a company called Vignette for 35% of their company. They quickly became the the fastest growing software company in history based on Story Server sales.

    In 1997 I was looking for CRM software for CNET and met Marc Benioff. In 1999 I became his partner, largest investor and second largest stock holder after only Marc himself. I went on Marc’s board which was the first outside board offer I had ever accepted.

    What these 2 things have in common is that in both cases I understood the problem space very well as someone in need who is both technical and a business person. I was therefore able to recognize the solution. Grandcentral.com, now Google Voice was started because I live in 2 cities and have 3 different cell phones, Skype, etc and could not manage my communication needs. We at Minorventures found Craig Walker at Yahoo and he solved my problem and created a great company at the same time.

    I have wanted for years for Swivel to solve “A” problem and not address many problems. If you solve 1 you can solve 2 and so on. Swivel is now being retooled to handle the massive amounts of data coming from atmosphir.com, a Minor Studios “game” which is growing rapidly. We have already integrated 3rd party data sources into a “mashable” data environment. I want living data that is not bound by a single data source or type. One final note is we want it to better leverage cloud services. I have believed all computing was heading this way since talking to Marc in 1997 and Swivel should be a paradigm for how its done.

    When its good enough for me I am glad to share it with everyone else and you can decide for yourself.

  16. Jonathan Avatar

    Creating watermarked charts is free; embedding clean charts on your blog costs $1 a month for first 1,000 hits, $30 for next 10,000, etc… Boom! Business model.

    I know it’s not that easy, but for reals, is it so hard to charge for a product up front? Because then you know which features matter and which ones don’t. The ones people pay for are the ones that matter. Develop those, tnx.

    Not mentioned here: I found Swivel.com (circa 2007) buggy and hard to use. And, yes, I work with statistics.

  17. Halsey Minor Avatar

    If anyone is still paying attention to this thread is their interest in our getting the service back up while we focus on solving SPECIFIC problems within minor ventures that will hopefully provide the insight needed to nail functionality, UI, use cases and reliability?

    Quite frankly I have spent 4 years investing in the concept and was a frustrated user myself. It did not live up to the general quality of services i would like to believe we produce. If people get value out of it I will bring it up publicly but it will be some time before I am doing so with pride. Having watched some great people take a swing at it I think I know what needs to be done and data visualization is not as important as people here think. Data visualization is a fraction of how we communicate information or even data.

  18. Tod Landis Avatar

    Hello, I visited your site several times after hearing about it from someone at an SF MySQL Meetup. It had a lot of promise. A desktop application, or a cloud-based application that worked like a desktop application could supply the context you were missing. I hope some of you will take a look at Entrance, http://dbentrance.com/ to see our thinking. Two features are relevant: Entrance has an import dialog that remembers import urls and their corresponding settings, making it easy to repeat imports easily. This is useful for things like CPI, which update monthly. It also provides a “plugable” arrchitecture for adding import tools to its “X” menu. Get in touch if you would like to see a demo, brain storm business models etc., I’m very interested in hooking up Entrance to data sources.
    – Tod Landis, Boulder Creek, CA, tod@dbentrance.com

  19. Kumar Avatar

    I liked their idea, i wonder how come it happened.
    It must be simple than it was.
    I tried with few data set, but fed-up with working on process.

    They may need more developers to make it bit simple..

    I added my small bit.. :)

  20. Anthony Calabrese Avatar
    Anthony Calabrese

    So wait… Swivel is coming back? I would be more than happy to beta test and trouble shoot their new product, but I don’t think that it is fair to the users that Swivel just closed up shop, turned off the lights and killed thousands of live visualizations.

    The trust is gone and Swivel has to earn it back in my opinion.

    Data visualization has a bright future, but only if the wider community remains engaged we have the right tools to tell our stories.

  21. Bear Avatar

    “build a product around some specific kind of data and use Data Visualization in the product, but don’t ‘sell’ it as a Data Visualization tool. For example, if you’re a project manager, and you want to measure performance of your team and projects, your more likely to have a ‘project management problem’ than a ‘Data Visualization problem’. So a project management tool will more likely fit the problem than a general purpose Data Visualization tool.”

    Yes!!

  22. Martin Trgina Avatar

    Hi,

    I think Swivel was nice. In my opinion there was really no sufficient business model. I mean pure on-line BM is always hard to run without real world sales people. Recently we have also started public data analysis market Mercato.belladati.com . http://Mercato brings data to journalists and people interested in data in general. However Mercato is based upon business from our cloud BI http://www.belladati.com.

  23. Anonymous Avatar
    Anonymous

    great

  24. ataur amini Avatar

    RailsXLS is a plugin to generate Excel files from a .rxls view. It uses a Java Bridge to talk to the Jakarta POI library.[arl=http://shakibmdasif.wordpress.com/]shakibMDAsif[/aul] I’ve used this plugin in a few projects and it works great for generating .xls files.

  25. dipu mazum Avatar
    dipu mazum

    i think it is a beautiful site.I have wanted for years for Swivel to solve “A” problem and not address many problems. If you solve 1 you can solve 2 and so on.