Perspectives on entrepreneurship, startups and venture capital from K9 Ventures.

The investment that didn’t happen

Tweet about this on TwitterShare on FacebookShare on LinkedInShare on Google+Email this to someone

Note: All information contained in this post is based on my best understanding and perception of what transpired. I have confirmed with the founders that none of the information contained herein is deemed confidential and is therefore fair game for me to share in this post.

In April 2008, I was introduced to Ugmode, Inc. (thanks to my advisor Terry Winograd). The founders of the company — Arlo Faria and AJ Shankar — were PhD students in Computer Science at UC Berkeley. Arlo and AJ had built a stunning new way of doing product discovery for soft goods. I define soft goods as the category of goods where you’re shopping based on the appearance of the goods (for example, shoes, clothing, sunglasses, watches, jewelery and furniture) rather than the specifications (most consumer electronics fall into this category). So far, almost every shopping website shows products either in an endless list of “Next Page” links (hint hint Amazon — you have so much room for improvement!), or by using keyword search, which doesn’t really cut it. Arlo and AJ had devised a way to use shape and color as the two axes that helped to discover soft goods.

Their site — — started out with shoes. When you arrived on the site you saw a diverse set of shoes. You picked the one that was closest to what you wanted (for example beach flip flops vs. dress shoes) and then Modista used that shoe as the seed for showing you a new collection of shoes, sorted by color on one axes and shape on the other. The site automatically adjusted to any display size (anywhere from the iPhone to a 26″ widescreen display) to show you the maximum number of shoes possible. Each time you clicked on which one you liked and the site refined its selection of shoes based on your input.

It created a fabulous experience for browsing soft good inventory based on what you’re visually looking for. In the user’s mind he or she would be thinking, I want a red shoe, with high heels, but you don’t need to translate that into keywords, you just click on what you like till you get closer and closer to what you like. Progressive refinement at its best.

On the back-end Modista was doing some heavy duty computer vision and image similarity comparisons to produce these results in near real-time. As a two person team Arlo and AJ built out the whole site and the back-end technology — and built it to scale — at one point they did a A/B test on Zappos and were easily able to handle all the traffic coming to the Zappos test site while being able to show results from the enormous Zappos inventory.

Modista was the perfect fit for K9. It was a team of two sharp Computer Science PhDs, they had core new technology in the form of the algorithms they ran in the back-end, and also core new technology when it came to the display of information. Their business model was to provide retailers with a new interface for shopping for soft goods — something that hasn’t changed a whole lot since Web 1.0. Modista had built Shopping 2.0.

I closely watched the company’s progress over several months and was in awe of the quality of the founders and their execution. On multiple occasions I expressed my interest to invest in the company. The first time I asked them what they needed, they said they just need to crank out code. And crank it out they did. On future occasions, whenever I expressed an interest to invest, the founders came back and told me that they had an acquisition offer on the table! This happened not just once, but multiple times. Every big company in the online retail space wanted to buy them at some point along the way.

One occurrence of this was in Q1/Q2 2009. Modista was in talks for being acquired by a major player in the retail space – While Zappos was doing diligence on Modista, the issue was brought up that one of other companies in this space —, founded by Munjal Shah — had filed for some IP in this area. Modista had been forewarned that had a “history of being litigious”. While I cannot be certain of what the actual reasons were, my speculation is that Zappos didn’t want to rock the boat by going through with the Modista acquisition, because concurrently with their discussions to acquire Modista, they themselves were engaged in discussions to be bought by the 800-pound gorilla in the space —

In August/September 2009, the founders and I agreed to work together to raise a round of financing for the company. I helped introduce the company to various angels and lead the effort to form a syndicate for their fund-raising round. We pulled together about $600K of commitments and interest, for a $500K-target financing round.

As I mentioned, Modista’s closest big competitor in the space was was already aware of Modista, and, (I believe) was also aware of the fact that Arlo and AJ were out trying to raise money. The day before we were supposed to sign the term sheet for the investment, sued Ugmode (the parent company of for patent infringement. Apparently, had a patent (‘610 patent) issued about two months before in the broad area of using computer vision for product search.

In my opinion, this is a patent that should never have issued in the first place since there is a lot of prior art in the space, including a lot of academic literature that should have (again, in my personal opinion) invalidated the patent to begin with.

To Arlo and AJ’s credit, as soon as they got served, they called to let me know what was going on. The lawsuit completely killed the financing prospects for Ugmode. The estimate for resolving a patent infringement dispute was about $1M in legal costs and the process could potentially take at least 2 years. Regardless of how much I liked the founders of the company, I could not in good conscience invest my LPs’ money in a company where a significant amount of the proceeds of the financing, if not all, would go straight to lawyers. The reaction from all the other committed investors was also appropriate and unanimous: “We love these guys, and it sucks to see this happen to them, but we can’t sign up for a lawsuit.”

At this stage, the team had had no prior financing. They had a small amount of cash in the bank from having won two business plan competitions that they used to finance the company to date. To put in bluntly, Modista didn’t even have the cash to be able to hire a patent litigator to defend them in this frivolous (in my opinion) lawsuit brought about by a company to simply try and reinforce its patent position against bigger players.

Modista had only had modest revenue to date (~5K) and had by no means gotten to the point where it was having an impact on the revenues of However, it was clear to me, and probably to as well, that even as a two-person team Arlo and AJ were technically far superior to anything could ever produce and hence were in the long term probably a threat to (Note: never asked for any damages in the lawsuit and only wanted an injunction against Modista)

Arlo and AJ are way too nice to go around slinging mud at anybody. And it’s been with their guidance that I have refrained from telling this story publicly before. They prepared a fairly long dossier of information which showed that the patent should have never been issued by the PTO and also dissected the patent sufficiently to show that even if the patent was upheld, they weren’t doing exactly what was in the patent anyway. However, even though as Computer Scientists they could examine and explain this to any other person who was a Computer Scientist, explaining it to a judge and to a jury would probably be a different ballgame altogether. That falls into the realm of hiring and paying for high priced litigators who can twist and turn words for an audience that isn’t an expert in the subject.

Apparently in a meeting with’s founder and CEO, Munjal Shah, they were essentially told that they should give up their own startup, and just come work for I don’t know about you, but as an entrepreneur if someone threatened me that way I’d tell them exactly where to put their words and their smart-ass idea.

Apparently a corporation can only be represented by an attorney in a litigation and cannot be represented by the founders or officers, even if the company cannot afford counsel. Long story short, the lawsuit torpedoed Modista’s financing. And since they didn’t have the money to hire lawyers to defend themselves, they had to accept a default judgment against the company, which forced them to shutdown

Modista shut down notice

A few months later, when I heard that Google was acquiring, I was only more disappointed. Google so far has played nice with startups and has had a positive impact on the startup eco-system (especially for exits, less so for hiring!) Google now owns the patent that caused to be shut down. I sincerely hope the folks at Google read this post, so that they know what they bought.

The Valley thrives on innovation and creativity. Arlo and AJ demonstrated their innovation and their creativity in what they created with Modista. It is really sad to see two sharp young entrepreneurs spend more of their time learning about the intricacies of the legal process, rather than getting to spend time on building their product and their company. It is even more sad to see companies such as that prey on small young startup founders before they’ve even gotten out of the gate.

I’m all for competition and for a fair fight. Fight on the basis of a better product, a better user experience, fight on pricing, fight on things that show that you are better. But using your big VC dollars to put fledging startups out of business, that’s just in bad taste. To that effect, it’s probably good to also mention that was backed by First Round CapitalBlueRun VenturesBay PartnersLeapfrog VenturesMenlo Ventures and Crosslink Capital to the tune of $47.3M, according to Crunchbase.

The purpose of this post isn’t to debate the pros and cons of software patents. If you’re interested in that discussion, my advisor and mentor, Brad Feld has written extensively about software patents on his blog. My friend Daniel Tunkelang has also authored several blog posts on this topic on his blog, some of which are inspired by (and also mention) Modista. To summarize that discussion, the system is broken at many levels. Starting from the overburdened PTO that issues these software patents all the way to the complexity of IP litigation which makes it prohibitively expensive to defend against unwarranted lawsuits. has been assimilated by the big G, so we can certainly hope that they won’t be out to get other fledgling startups (and yes, I’ve seen many more startups using computer vision based techniques — something wanted the world to believe they had a monopoly on). However, the IP climate remains unchanged, and it’s likely that other villains will continue to exploit similar situations.

The Valley is a fascinating place — you hear stories about it every day. But not all the stories in the Valley get told. I didn’t want the story of Modista going untold and so I’m taking it upon myself to share their story — from my perspective.

We all like a story that has a good ending. This one didn’t end well. But, having worked closely with Arlo and AJ through the thick and thin of this journey I have nothing but the utmost regard for them as entrepreneurs. I told them then that “Whatever you do next, I’m ready to back you anytime.” They’ve both since moved on. Arlo is back at Berkeley wrapping up his PhD, and AJ is now working on a new startup, which I’m proud to be already be an investor in.

So here’s a toast to — the site that built the best new shopping technology I’d seen in a while, but didn’t survive.

You can follow me on Twitter at @ManuKumar, or, follow @K9Ventures for just the K9 Ventures related tweets.

  • Vivek

    Can someone update how is back? Whats the end story?

    • The site is now back online, but AJ and I will leave it in its original state, circa 2009. I’d like to thank Google’s lawyers for their help. You can still use the site, and the inventory is current (thanks to the great Zappos/6pm API), but we’re not making money from it. Thanks everyone for your nice comments!

  • Emilyellen1

    Can someone fill me in as to what else happened with Modista? Going to the site now, it is up and running.

    • manukumar

      @Emilyellen1 I’ll ask Arlo Faria to respond to you.

      • Emilyellen1

        @manukumar  Arlo Faria  Thank you. It would be great to know the rest of the story and what Arlo and AJ are up to now.

  • Pingback: Patents: What are some good patent stories? - Quora()

  • Pingback: Munjal shah How is a scammer like Munjal Shah allowed to roam freely? - Quora()

  • Sorry to see a promising start up close down so early in the game!

    Start ups are less productive…trying to stay clear of patent mine fields. Very disappointing!

    Thanks for the story. Good luck with the new venture!

  • max.grant


    You can’t blame either,

    Yes, we can. Every company has a choice as to how they behave. It’s a lame copout to hide behind “we’re just doing the best for our shareholders.” The execs of could have also gone on a killing spree in order to eliminate competition, or accused the Modista guys of garden-variety theiving, or planted drugs on their premises and directed the police to them. None of these things would have been legal for to do, but in the short term it would have been “the best” for their “shareholders.”

    Unethical behavior is still unethical, even if it’s not illegal.’s patent injunction was based on a falsehood. The fact that they got away with it does not excuse it. The fact that they were purportedly doing it “for the shareholders” does not excuse it. It’s flat-out inexcusable behavior.

  • joe.f.kelleher

    The fact you can’t represent yourself in a patent dispute sounds like yet another broken part of the patent system. In a criminal case, you can either represent yourself or have a qualified lawyer represent you, but you can’t get your friend who studied a bit of law to help you out with represenation. Supposedly this is to protect defendants from employing unqualified representatives and jeopardising their cases. But for a patent claim against a company: i) a comapny can and should be trusted to appoint the right person to represent it, regardless of formal qualifications, which may mean the founders represent tehmselves, and ii) a domain-specific technologist who learns about patent law in the course of their work may actually be better qualified to defend against a bogus patent claim than a lawyer who chooses to specialise in patents generally.

    I’d be interested to hear if the situation is different in other countries. I’m from the UK, where the ruling party has (surprisingly imo) made some recent noises about ensuring intellectual property law actually helps economic growth.

  • ayemoah

    What do you guys think of Paul Graham’s Patent Pledge Project?

    Fundamentally, a company that intends to use its patents offensively won’t pledge in the first place. So it may not have prevented Modista but it’s a good place to start if we can get a significant number of companies to commit to it.

  • nickziv

    This illustrates the damage that the patent system causes to society and the economy by killing innovation in the cradle. Ironic, seeing as how the system was initially designed to protect and foster innovation. It seems that the patent system is beyond reform, considering the overwhelming number of patent applications, the granting of patents that are clearly prior art (which is an embarrassing testament to undeniable incompetence on the part of the PTO), and the astronomical cost of defending against a bogus, prior art patent. The patent system (at least in the field of software), must be dismantled, as that is the only efficient way end this parasitic drain on society quickly, and let innovators be rewarded (instead of being punished) for daring to innovate.

  • laughinghan

    What pisses me off the most about this is that the consumers are the real losers, here. What I’m hearing is, it is perfectly possible for all our lives to be better next time we search online for soft goods, but the system fucked up and killed this innovation. This is the opposite of everything the system is supposed to do.

    If they don’t plan to use this tech on their next startup, why don’t they make it open source? Once something has been invented, it cannot be un-invented. Now that we know this is possible, I’m sure I am not the only one who, tonight, is going to start trying to replicate what they did. If they’d open source it, it would save someone else in the world a lot of wasted effort.

  • DarrylSiry

    @manukumar +1 “Apparently in a meeting with’s founder and CEO, Munjal Shah, they were essentially told that they should give up their own startup, and just come work for I don’t know about you, but as an entrepreneur if someone threatened me that way I’d tell them exactly where to put their words and their smart-ass idea.”

    Of all the aspects of the story that part reveals the most about the character of the folks. Let me translate – “Not only am I going to kill your dream, but I’d also like you to be my bitch. How about that?”

  • Pingback: My favorite stories that happened while I was taking finals « Daniel Miller . Law . Tech . Games . Whatever()

  • chrisyeh

    @carterac I wouldn’t engage in such behavior, but there are plenty of VCs who would view a decision to not block a competitor’s funding as a major strategic mistake. None of us have any idea what kind of shareholder pressure Like might have been under at that point.

  • CharlieS

    @chrisyeh @manukumarReally?! Since when is it OK to have one of your portfolio companies engage in unethical behavior to maximize the value of your investment?

  • CharlieS

    Thank you for letting this story see the light of day. When Riya/Like launched with a great deal of effort, I remember checking out Munjal Shah’s blog and actually being impressed by the guy. I no longer hold that opinion, and I hope that since the Valley is a small place, people remember your story when considering doing business with him.

  • carterac

    @chrisyeh @manukumar No-one is debating the effectiveness of the strategy, but rather its unethical nature. Does shareholder value always trump ethics?

  • nikcub

    Sad story. I can see the argument about fiduciary duty, but then at the same time there are companies that hold patents only for the purpose of defense, rather than as a mechanism to undermine potential competitors.

    There are funds and companies that invest in startups that have this specific problem. There must be a way to get in touch with potential investors who would take the opportunity to challenge the patent in litigation. Patent hoarding and litigation in software is an industry unto itself.

    I can’t help but think that the round of financing could have been structured to cap legal expenses regarding the litigation to 80-130k. That would get you some way towards discovery and pre-trial, and towards negotiating a settlement. Most startups end up spending as much on legal fee’s regardless. A combination of a cap along with a law firm that would waive the remainder of fees until a later date for cash, or having the remainder of their fees convert at the same terms, could work. I hope there is a good patent litigator out there who hears this story and decides to help you out.

    The litigation would take years, and the company and product seem like a great acquisition target – since this type of technology will play a large part in the future of e-commerce. I hope a solution can be found.

  • danlucraft

    @chrisyeh @manukumar Just because the law gives you license to act like a Mafia thug, doesn’t mean you should.

  • Risk

    Thanks for sharing this story Manu. That really sucks about Modista. Was even a successful exit?

    I really feel like Amazon could easily pull this technology and the only reason they haven’t done it yet is because people don’t want to find products that way.

  • iptiam

    Couldn’t they have applied for and received a patent for what they were doing? from what you say, sounds like they definitely had some solid new tech?

  • dtunkelang

    @dineshbvadhia I can’t remember the terms under which IBM released those patents — but I wonder if one of them might have been the condition that you not sue IBM for patent infringement. Google recently bid $900M for Nortel’s patents for defensive purposes — see

    And even questionable patents can still have defensive value — a point I raised in my “practical rant”:

    In fact, Google’s announcement a few weeks after my blog post made me feel very prescient. 🙂

  • dineshbvadhia

    A few years ago IBM released a bunch of its patents into the public domain. It would be wonderful if Google did the same for this particular patent as it serves no one including Google. As far as I can tell, the patent doesn’t describe any technology eg. algorithms but a method for searching for merchandise using images. @dtunkelang

  • dineshbvadhia

    Looked at the prior art and I’ve seen similar stuff before (as I’ve been involved with CBIR for quite a few years). It is totally unbelievable that such a broad based patent can be granted. A few minutes of searching in the academic literature or Google would uncover a ton of prior art.

    I really feel for the Modista guys. Better get back to work – got to get my patent on a hammer into the PTO by next week before some other bugger invents a hammer.


  • dtunkelang

    @dineshbvadhia Good attorneys know how to convince a patent examiner that a claim is extremely narrow while in fact obtaining extremely (and often overly) broad coverage. Moreover, most patent examiners lack the specific domain knowledge to understand the true breadth of a claim. But it’s unrealistic to expect that our best domain experts will pursue careers as patent examiners. Did I mention that the system is broken?

  • dineshbvadhia

    Okay, understand you can’t comment.

    The idea that blanket coverage can be given in a technology area that is still in its infancy with many unresolved problems is beyond ridiculous.

    Fortunate for to have got away with it. I always thought the image recognition solution was pretty crap!

    @dtunkelang @dineshbvadhia

  • dtunkelang

    @dineshbvadhia Dinesh, I’m not in a position to make such an analysis and comment on it publicly. But you can look at the patent claims and some of the prior art yourself, as Manu as provided links in the post, and I believe you will arrive at the conclusion that Modista would have triumphed on the merits, given the opportunity. Also, while Google holds this and many other patents, Google has not, to the best of my (and Quora’s) knowledge, ever sued anyone for patent infringement. I hope that Google’s acquisition of means that no one will ever use this patent offensively again.

  • dineshbvadhia

    @daniel @manukumar Haven’t delved deeply into the’s patent but I find it hard to believe that it has such broad applicability. Consider, property purchases on realtor sites starting with the text query “San Francisco, Mission, 2 bedroom apartment” and the service would show N images of properties that match the query. Next, select one or more images (which may or may not have associated descriptive text) to refine the search. Continue until satisfied. There are many other product categories similar to this eg. industrial parts.

    Does have total patent coverage on all such applications?

  • nutanc

    Very well told Manu. Really unfortunate for Modista. If they are not using some of the technology in their current work, it would be great to just open source Modista. Put up the code on github or sourceforge. Lets see them sue the open source world 🙂

  • alain94040

    Thanks for sharing a great, sad, story.

  • anand venkataraman

    Wow Manu, a very nice and thoughtful post. As I have come to expect from you! Long live Modista (in spirit at least)

  • ayemoah

    @manukumar @ayemoah @stevenan Yea, it sounds too much of a stretch even in my imagination.

  • stevenan

    @manukumar @ayemoah @stevenan Interesting. It seems very odd that this legal system is dependent upon a set of very high priced skills that the average person apparently does not readily have. If you can’t pay for them, you can’t win, even in a possibly frivolous case like this. But I guess this is the type of thing that needs to be reformed in our patent system…

  • manukumar

    @ayemoah@stevenan Patent *litigation* is a rather specialized field. There are lots of people who can help draft (prosecute) patents, but not as many who can handle litigation. We did activate as much of the network as we could, but didn’t find anyone who would take this on without requiring cash payment. It is a significant investment of time and resources and also has some hard costs associated with it, so that is understandable. That said, several patent attorneys did generously give their time to help advise the company on what the options were — but signing up to be the attorney of record in a litigation case is a much higher bar.

    While it is common for plaintiff’s counsel to be contingency based, AFAIK that hardly ever happens for defendant’s counsel.

  • manukumar

    @stevenan This is exactly what Arlo and AJ wanted to do, but the legal system didn’t permit them to do it (and yes, they tried).

  • ayemoah

    @stevenan In the same line of thinking, could they have convinced a friend who has a law degree and taken the patent law exam to represent them in exchange for some % of equity? If they won, the company lived, the friend would have gotten some equity and the financing round could have happened. If they lost, well they lost and the lawyer friend got nothing.

    It’s probably quite a stretch (not everyone knows a friend who’s a patent lawyer who wants to help) and the probability of success is quite low. Multiple layers of ‘if’ are involved for this to end happily. I mean it has to be a close personal friend who wants them to succeed and who is willing to spend his personal time without a guarantee of being compensated for the effort? Many lawyers right out of school are getting furloughed in the recent years and in my alternate optimistic world, this could happen. 🙂

  • stevenan

    Really interesting story. I’m a CS PhD student myself, so let me ask a dumb question: What would be the harm in Arlo and AJ just stepping in to court representing themselves? Sure, chances are they will lose, but would it cost them anything to just give it a shot?

  • Pingback: The Story of a Startup | The Kindness of Strangers()

  • chrisyeh

    @manukumar What did wasn’t nice, but it was in the best interests of its shareholders. They viewed their patent as an asset and Modista as a threat. Killing Modista’s round (while reprehensible) reinforced the value of their asset and took out a competitor. I can think of plenty of investors who would be upset if a portfolio company passed up that opportunity.

  • manukumar

    Chris, If Modista was having an impact on’s revenues, then you can make the case for the responsibility to shareholders. As I said in the post, all for a fair fight, but to me this appeared more of a case of doing it because they could, and, also perhaps to build strength around the enforceability of the ‘610 patent. has gone on to sue other companies (SuperFish) using this patent. From what I’ve heard Google is still pursuing that litigation.

  • dtunkelang

    @chrisyeh I agree that the patent system is broken, but I’m not so quick to forgive companies that abuse its brokenness in the name of fiduciary responsibility. Ethics should come before shareholders. Or, as others might say, “Don’t be evil”.

  • chrisyeh

    Excellent story, though an upsetting outcome. You can’t blame either, since they too have a fiduciary responsibility to fulfill to their shareholders. The blame lies firmly with the broken patent system that allows these overly broad patents to issue to begin with.

  • AlexSherstinsky

    Manu, thank you for describing the entrepreneurial journey of Arlo and AJ so clearly. Best Regards, –Alex

  • ankur10

    Thanks Manu for sharing the story. This is a great example of how much our patent system is screwed!

  • dtunkelang

    Manu, thank you for telling this story in all of its gory detail. As a witness to this story, I was frustrated to see how Arlo and AJ never had a chance to win on the merits of their case, let alone on the merits of their technology. I can only hope that this story inspires others and drive needed software patent reform. And that Arlo and AJ both go on to do great things — as I am sure they are doing already.