A Startup A Day

For more information about the Y Combinator Challenge, check out this post first 

14. Tools for measurement. Now that so much happens on computers connected to networks, it’s possible to measure things we may not have realized we could. And there are some big problems that may be soluble if we can measure more. The most important of all is the defining flaw of large organizations: you can’t tell who the most productive people are. A small company is measured directly by the market. But once an organization gets big enough that people on in the interior are protected from market forces, politics starts to rule, instead of performance. An improvement of even a few percent in the ability to measure what actually happens in large organizations would have a huge impact on the world economy, and a startup that enabled it would be entitled to a cut.

My Idea – Coworker Karma

I follow a pretty simple process for coming up startup ideas for the Y Combinator Challenge series.  Each night before I go to  bed I read the next idea on the list on my mobile phone, think about it for a few minutes until I fall asleep, and when I wake up – poof!  There’s my idea.  It’s kind of a fun little trick that I’ve been doing for a while, and it’s usually pretty reliable.  However, this morning when I woke up…nothing.  Zip.  Nada.  I could practically hear the squeaking as the hamster ran around on his wheel.

I have a good theory as to why my technique didn’t work this time.  It’s because, for all intents and purposes, this idea is impossible.  At least, if you look at it holistically and come up with a model that would be consistently applicable to every worker in every role at every level.  And, you’d need it to be 100% accurate and able to be defended against any claims of bias (both legal or otherwise) by disgruntled employees.  Therefore, I’m going to change the approach a bit and only focus on one segment of this problem – after all, as Paul points out, even a small improvement can be a big win for companies.

My idea is based on a pretty simple hypothesis:

1. People generally have a pretty good idea who the high performers are within their organization
2. The more closely you work with someone, the more accurate your assessments of your co-workers are

So, my idea is a system that would allow for users to give “karma points” to their co-workers each time they do something helpful, and then to combine the karmic data with additional statistics to create detailed reports for the management team.  In order for employees to participate, the ability to give points would need to be extremely simple and integrate seamlessly with their everyday workflow.  The front end would be a simple Outlook plug-in that would add a little “give karma” button next to the name of the person who sent the Email.  If the Email contains something helpful to you, instead of writing back a meaningless “Thanks!!!!!11ONE”, you can instead click the karma button and give them a karmic point (and provide optional comments).

On the back end, the karma points would be tallied to create a karmic score for each employee, but it wouldn’t be a simple sum.  The system would be linked in with the company’s hierarchical structure (do most big companies have that, or is it an MS-only thing?), and would weigh karmic points received from immediate co-workers or direct managers above those received from outside the team.  This information could also be used to assign extra weighting from VIPs (i.e. the general manager or VP of an organization).  The points should also be normalized, so that a user who only gives out points for above-and-beyond work isn’t dominated by Bob in marketing who gives out a hundred karmic “boo-yahs!” on a daily basis. 

For managers, the system would produce a series of slick-looking reports that would aggregate the data from the past 3/6/12 months and present reports for their specific team or division.  You could drill into individuals to view comments or view historical performance to see how they are tracking over time.   Ideally, the manager would this data to augment (rather than replace) traditional performance assessment tools – this would work especially well for companies that “grade on a curve” to encourage teamwork instead of a competitive me-vs-you environment.  Also, since there is no concept of negative karma or downvoting, the system would only provide additional data point to recognize high performers as opposed to weeding out workers who are under-performing.

What do you guys think?  Any products like this out there today?  Another good take on this idea that’s already in the market is a system from Traverse Corp. They were part of our “A Startup A Week” series – great guys and lots of smart ideas on how to tackle this big problem.

For more information about the Y Combinator Challenge, check out this post first 

13. Online learning. US schools are often bad. A lot of parents realize it, and would be interested in ways for their kids to learn more. Till recently, schools, like newspapers, had geographical monopolies. But the web changes that. How can you teach kids now that you can reach them through the web? The possible answers are a lot more interesting than just putting books online.

One route would be to start with test prep services, for which there’s already demand, and then expand into teaching kids more than just how to score high on tests. Another would be to start with games and gradually make them more thoughtful. Another, particularly for younger kids, would be to let them learn by watching one another (anonymously) solve problems.

My Idea – SuperMemo High

There was a great article in Wired a few months back that profiled a Polish man named Piotr Wozniak (no relation) who created a software program that uses a memorization technique called the spacing effect.  The basic idea is that memory fades in a surprisingly predictable pattern, and if you can reinforce the initial learning at very specific intervals, you can dramatically increase memory retention rate.  Here’s a diagram that explains this a little better (courtesy of Wired):

Of course, school teaches us much more than just rote memorization.  You learn problem solving, teamwork, connections between events, social skills, etc..  So while today’s idea is centered around using the SuperMemo technique, any real attempts at providing a full online learning system would need to be fairly broad.

Here’s a take on how this could be implemented.  To start, each day’s lessons would be broken down into 10 minute sections – each hour would have 5 segments focused on traditional subjects (math, science, geography, etc..) and one free segment where users could take a quick break and/or socialize with each other via webcams, public chat rooms, or private IMs.  The learning segments would be compiled by experienced educators and could be presented either as text, video, interactive simulations, or even educational mini-games. 

For information that needs to be memorized, the lessons would be strategically presented at the proper intervals to align with the principles of the spacing effect.  Also, as students work through their lessons, the system would adjust the difficulty to ensure that each student is working at a challenging (but not frustrating) pace.  In addition, while most segments would be individually presented to the student based on their current pace and spacing, certain segments would bring together 2-4 students to work on a problem together.  Ideally, the students would not have more than 20 “classmates” that they interact with during the team sessions or the breaks, which would allow longer-term friendships to form as they would in a real-world classroom environment.

What do you guys think of this idea?  Believe it or not, I actually spent a year in a middle school computer lab as a teacher.  I can tell you from firsthand experience that computer-based learning systems are very expensive and have lots of room for improvement.  As always, feel free to post any comments below.

For more information about the Y Combinator Challenge, check out this post first 

12. Fix advertising. Advertising could be made much better if it tried to please its audience, instead of treating them like victims who deserve x amount of abuse in return for whatever free site they’re getting. It doesn’t work anyway; audiences learn to tune out boring ads, no matter how loud they shout.

What we have now is basically print and TV advertising translated to the web. The right answer will probably look very different. It might not even seem like advertising, by current standards. So the way to approach this problem is probably to start over from scratch: to think what the goal of advertising is, and ask how to do that using the new ingredients technology gives us. Probably the new answers exist already, in some early form that will only later be recognized as the replacement for traditional advertising.

Bonus points if you can invent new forms of advertising whose effects are measurable, above all in sales.

My Idea – Woot Ads or Judy’s Ads

Let’s mix things up a bit today.  I came up with two ideas for this idea last night and I’ve been debating which one to post.  So instead of picking one, let’s try something new.  I’m going to post both ideas, and then in the comments y’all can vote for your favorite one.  If you choose the idea that gets the most votes within the next 24 hours, you’ll win the grand prize – your name etched into the moon with my giant laser.  What?  Inside joke, nevermind…

OK, I’m gonna start with WootAds.  Hopefully you’ve all basked in the glory of that which is Woot.  If you haven’t, the concept is pretty simple.  Take a single item and sell it for a limited time at a ridiculously low price.  Only one item is available for sale at any given time, and once it sells out it’s gone for good.  Along with being a great opportunity for deals, it is, as the kids say, hella fun.  This is nothing new, of course – just a twist on the old school “loss leader” strategy that brick-and-morters like Best Buy have been using for a while (sell CDs at a loss to bring customers to the store, and then make it up on ridiculously high margins for big ticket items like flat-screen TVs) 

My idea for a new style of ads is based on this same simple premise.  Instead of static text or an image that you would click on to go to a site, the ad would instead feature one heavily-discounted item that is only available for a limited time.  The ads would be fairly simple, consisting of a picture of the item, a one sentence description, and a clock that would count down the number of minutes left until the item would no longer be available at that price.

What are the motivations on both sides for this one?  For consumers, it would elevate a boring old ad into a fun way to potentially find an amazing deal on something they may not have even known they wanted.  For retailers, it would give them a way to offload overstocked items to impulse buyers and to bring new customers into their online stores who already have all intentions of going through the checkout process.  Also, this is extremely easy to measure the ROI – the retailer pays based on the number of featured items that were sold.

The next idea I’m going to call Judy’s Ads.  It’s basically Judy’s Book for ads, and features user reviews of a product directly inline with the ads.  The goal of this approach is to get around the key issue of truth in advertising.  People are much more open to trusting other people’s opinions about a product than a carefully crafted message from the businesses selling the product.  So why not tap into that same basic concept within the ads themselves? 

Here’s how this could work.  Each ad would have two sections – a top section with a simple AdWords style description and link to the product/service, and a bottom section with two options for rating the product, either thumbs up or thumbs down.  After rating the ad, the bottom of the ad would reveal the overall rating of the product by other users, and would provide options to read or post comments. 

Why would this work?  First, this would discourage advertisers who are advertising crappy products or shoddy services, which would lead to better quality services being advertised, which would instill more trust in consumers which would result in higher click-through rates and so on.  Second, because the ads would be interactive, users would pay more attention to them.  Also, since the ads are already using a format that companies have researched and tested (standard AdWords format), they could take advantage of their previous investments and simply repost their same AdWords ads.  Finally, because the users would be encouraged to click and rate the ads they saw on each page, the system would be able to optimize the ads for the most active cross-section of the site’s user base, which is the audience that the ads would ideally be targeting.

All right – which one is it going to be?  WootAds?  Judy’s Ads?  Cast your vote for your favorite idea in the comments below.

For more information about the Y Combinator Challenge, check out this post first 

11. Web Office apps. We’re interested in funding anyone competing with Microsoft desktop software. Obviously this is a rich market, considering how much Microsoft makes from it. A startup that made a tenth as much would be very happy. And a startup that takes on such a project will be helped along by Microsoft itself, who between their increasingly bureaucratic culture and their desire to protect existing desktop revenues will probably do a bad job of building web-based Office variants themselves. Before you try to start a startup doing this, however, you should be prepared to explain why existing web-based Office alternatives haven’t taken the world by storm, and how you’re going to beat that.

My Idea – OfficeHours

Considering that I’ve been happily employed at MS for the past three years, I’ve only got two words to describe this post: Awk. Ward.    But, a challenge is a challenge, so here goes.

If my idea had a thesis statement, it would go a little something like this: to gain share with an online office suite, you should look at The Facebook and follow the same pattern that Mark Z and the gang did when building their little empire.  To begin, target the college crowd in a very specific way.  Only allow the site to be accessed by students, and only allow them to access the section that is focused on their school.  Why college students?

1. College kids have no money.  And even if their parents gave them the $50 it takes to buy the discounted student edition of Office, 90% would rather spend it on beer.
2. College kids are surrounded by always-on, high speed Internet connections.
3. College kids have grown up using Internet-based applications and are not afraid to use them.
4. College kids don’t worry about privacy and don’t have to abide by any government or tax regulations around data privacy or accountability.
5. College kids often work in groups and have a need to share information.
6. College kids are generally already connected via some social network (probably Facebook).
7. College kids like to experiment and try new things.

In other words, they are perfect candidates for free, online Office-like apps.  Here’s how this could work.  Students would first log onto the site using their Facebook credentials (thank you, Facebook Connect!) and would be taken to a very cool-looking home page customized with the school’s logo.  First-time users would then be prompted to choose which courses they are currently taking (i.e. CHEM100, ENG425, etc..) to further customize their experience.  Once their schedules are set, they would have several options.

First, users could choose to create a new “Word”, “Excel”, or “PowerPoint” document.  Each of these standard office apps would have just the basic, most useful features, and would also allow users to invite students from their study groups (via either IM integration or Facebook messaging) to work along with them in real-time on the assignment.  Invited users could edit the copy directly or insert “thought-bubble” style comments anyplace within the text.  To enable flirting enhanced collaboration, the sidebar of the app would be configured to feature multiple real-time streaming video chats.  In addition, users would see names and Facebook profile photos of any members of their classes who are currently online, and could invite them in for an impromptu study session or to send them a quick question via IM.

While users would have the option to change the privacy settings on their documents, the default would be set to “only show to friends in my classes”.  When users would log into the site, they would see a news feed of all of the documents that were recently created by their friends, and could browse through them to catch up on any notes that they missed, validate answers to assignments they’ve already done, etc..  To help combat cheating, the app would keep track of every person who viewed each document (and for how long) and give the document creator the ability to block any freeloaders from any future documents.  You could go even further and print out a unique URL at the bottom of each printout of the documents, which the professor could use to view a page listing all the stats of the various contributors and viewers of that document.

What do you guys think about this one?  Any college students out there that would use something like this?  I’d love to hear any thoughts on this one in the comments below.

Hi, my name is Charles Ju and I’m answering Kevin’s call for help. I am also an unabashed fan of Y Combinator (submitting a plan this year), and thought that I’d share some of my ideas with the community.
Please visit my blog at www.charlesju.com. If you want feedback on your ideas, a review of your startup, or just to meet a new friend feel free to e-mail me at charlesju[at]gmail[dot]com.

10. Auctions. They are a major foundation of internet revenue, eBay posted $7.5 billion USD in revenue last year. While this is not all from auctions, I think it’s fair to say that a large portion of it is. With the inclusion of sites like Craigslist and FB marketplace, it becomes clear that finding a new way to monetize this market will be enormously profitable.

My Idea – Mobile Auction Tweets
The pain with auctions is that it is not tailored for the casual user. If you want to auction off an item on eBay, for example, you need to take a picture of it, find a description, and make the page attractive. This takes too long.

My idea is to create a mobile application, preferably starting on the iPhone to streamline the auction process. Similar to the concept of microblogging, auctions need to be simplified. All it will take to create an auction using my application would be a picture from the camera of the iPhone (or video), a short description, starting bid, and duration of the auction. This application will exponentially decrease the amount of time it takes to auction off casual items. Ideally this application will be able to plug into different auction sites; Facebook Marketplace, Craigslist, eBay, etc.

There are a couple of benefits to this business model:
1. There is an actual business model. Each user will be a paying user and you won’t have to fight the constant battle to make sure your ads are paying more than your server costs. By selling an actual product with a price you’ll be cash flow positive from day 1.
2. Mobile auctions are already a big thing in Japan. So this model has been proven to work. http://www.dena.jp/en/work/mobile.html. Having a proof of concept in a foreign country will greatly decrease the inherent risk in starting the venture.
3. Avoids the problem of critical mass. Normally when you start any site that depends on the interaction of users you’ll have the critical mass problem. You need users to get people to use your site but no one will use your site until you get users. By tapping into existing networks like eBay and Craigslist, you’ll already have a full set of users ready to go.

Some risks off the top of my head to look out for are that big auction sites will not allow your application to use their API, a big auction site might take your idea and do it themselves, and there will be no demand for casual auctions.

Thanks to Charles for posting this great idea – feel free to post any comments below.  Also, be sure to check out Charles’ blog, he’s new to the blogosphere but already has some pretty good stuff.  If you have an idea you’d like to post, go ahead shoot me a mail at kleneway@hotmail.com.