We’re probably within 6-8 weeks of having our Genlighten private beta ready for interested providers to try out (though everything always takes longer than we’d like).  One of the decisions we need to make in the next few days has to do with where Genlighten will be hosted long-term.

Up until now, the alpha version of the site has been built in PHP and hosted at MidPhase on a “shared” basis.  This has been a cheap option while we’re just trying things out, but it’s time to get more serious.

The beta version of Genlighten is being built in Ruby on Rails (RoR) and will need a more sophisticated (and reliable) dedicated hosting environment.  Ideally, our new host would have startup-friendly pricing that would let us pay for only the storage capacity and bandwidth we need, then let us scale the amount we pay as traffic (hopefully!) grows over time.  Another criterion, since we can’t yet afford to hire an engineer who focuses her/his time specifically on server setup and management, is that the host offer a fairly high level of support, quick response in the event of a problem, and robust backup options.

We’re focusing on two main candidates right now:  Joyent and Engine Yard.

Joyent

Joyent is located in Marin County, CA and has been around for several years.  They’re backed by Peter Thiel, former CEO of PayPal, currently president of Clarium Capital Management and managing partner at Founders Fund (the VC firm that invested in Geni.com).

Joyent provides hosting for both PHP and RoR applications.  So for example they have offered free PHP hosting for developers of Facebook applications.  They’ve also provided RoR hosting for Twitter, though that relationship ended amicably earlier this year.  Both of these examples speak to Joyent’s ability to help web apps scale to large numbers of users.  Here at Genlighten, we hope we’ll need that kind of scale someday!

Joyent offers what they call a 4 GB Accelerator with 4 GB of RAM, 50 GB of storage, 1 CPU core and 10 TB of data transfer for $500/month and a one-time setup fee of $500.  That comes with a relatively modest level of support (including help if the site becomes inaccessible or slow).

Engine Yard

Engine Yard has been called the Ferrari of RoR hosting.  They’re focused exclusively on Rails.  Investors include Benchmark Capital, New Enterprise Associates and Amazon.com.  Engine Yard enjoys an excellent reputation in the Rails community.  They’re known for hosting both Github and Lighthouse, key developer tools for version control and bug tracking.

They offer a higher level of “managed support” — even for their entry-level monthly offering –  than Joyent does (or at least it appears so from their website).  You pay for that, naturally.  It’s tough to make an apples-to-apples comparison with Joyent, but I’ll give it a try.

Engine Yard recommends early-stage web developers start out with three slices — two for production (including one for backup) and one staging slice.  That costs about $1,050 per month (plus about $800 for initial setup), and includes 768 MB of RAM, a dedicated processor and 45 GB of storage.  Note that this is much less RAM than Joyent includes.  It’s not clear to me how much bandwidth we get for that price.

What makes this attractive are the support features.  Even at the entry-level we’re looking at, you get 24/7 support, database backups, Github and Lighthouse subscriptions, and shared load balancing.  From what we can tell, a lot less of our CTO’s time would need to be used for server maintenance and database management with Engine Yard’s service.

Slicehost

There’s another option we might consider if either Joyent or Engine Yard look like they’re going to be too expensive for our needs at this stage:  Slicehost.  Their service is at the opposite end of the spectrum from Engine Yard’s:  a quality service, but one designed for those who can handle server maintenance, software installation and downtime recovery themselves without the need for extensive support.

In return for a lot more of our CTO’s hands-on involvement, we could get by with a much lower monthly fee:  somewhere between $70-$250/month for a 1 GB or 4 GB slice, with 40-100 GB of storage and 400-1,600 GB of bandwidth.  Since we would do the setup ourselves, there would be no setup fees.

Conclusion

The hosting decision is an important but difficult one for us.  We want a hosting provider who can support as we grow but not charge us for services we don’t need yet.  On the other hand, we want a partner who will handle back-end server maintenance tasks as we scale so we don’t have to rush out and hire an engineer the minute traffic spikes due to a favorable review by a genealogy blogger.

In a recent presentation to entrepreneurs on how to survive the economic downturn, the famous VC John Doerr recommended:

“Negotiate. Negotiate with all your supplies and vendors, get more favorable payment terms.”

Our hope is that we can find a high-quality hosting company that can become a long-term partner, but that will consider offering us favorable terms while we’re still building our early customer base.  If you have any suggestions or recommendations, please let us know.

Scary Times

It’s been sobering to note the impact of the current economic uncertainty on web-oriented startups.  The tech blogosphere has been rife this past week with announcements of:

• Management restructuring
• Layoffs
• Dramatic efforts to conserve cash
• Websites closing down completely.

So how does Genlighten plan to weather the downturn?

Our Strategy

To begin with, since we’re completely self-funded, we’re not worrying about reassuring anxious VCs or angel investors.  We don’t have any non-founder employees (and we’re not paying ourselves) so no one’s going to get laid off.  We’ve got enough cash on hand for at least a year of operations at our current burn rate.  And we’ve all kept our day jobs, so our personal finances are in reasonable shape.

We are being a bit more selective about our attendance at genealogy conferences.  We had planned to be at the Who Do You Think You Are Live 2009 show in London this coming February.  But with airfare, lodging, exhibit fees and utilities for that show running about 10x of expenses for other major conferences, we decided that we just couldn’t afford it while we’re still in pre-revenue mode.  We do plan to be at FGS, NGS, IAJGS and the various shows put on by Family History Expos next year.

Heads Down, Moving Ahead, Preparing for Launch

Beyond that, we’re basically designing, wireframing, and coding like crazy — just like we were before the Wall Street turmoil began.  Our focus is still on building out the initial set of features for the launch of Genlighten’s public beta early next year.  Recently, we’ve been working on:

• streamlining the lookup offering process
• simplifying the checkout procedure
• incorporating a lookup activity feed in our user dashboard, and
• building out the design of our user profile pages.

Striking the right balance between optimum functionality at launch, sufficiently broad geographical coverage of our provider network, and the desire to simply “get something up and running” has been tough for us.  That and our transition from PHP to Rails have been the main reasons that our launch has been delayed thus far.  [That and the "keeping our day jobs" strategy.]  We know a number of you who we’ve met over the last year are anxious to see us finally go live.  Be assured that we’re eagerly anticipating that day as well, both so we can keep our commitment to you and so we can begin to bring in some revenue.

Building Our Provider Network

In parallel to our development efforts, we’re continuing to actively recruit potential lookup providers interested in offering feedback and participating in testing of the site prior to launch.  If you’d like to sign up for access to our private beta, you can do that here.  We’ve been highly gratified by the response we’ve gotten so far.

Staying Optimistic

Amid the gloomy prognostications of a potential startup apocalypse,  it’s been gratifying to hear a few voices quietly urging optimism, perserverance and long-term perspective.  In particular, entrepreneurial pundits are advising startups that the best way to survive in tough economic times is to focus on saving people money and helping them make some of their own.  Fortunately, that’s exactly what Genlighten is designed to do — for genealogists everywhere.

I’m here at the FGS Conference in Philadelphia looking out from our booth on the exhibit hall, which is buzzing with conference-goers.  We’d originally hoped to announce our beta release at this event, but as with many software projects, our milestone estimates turned out to be optimistic, and we’ve had to move our anticipated launch date back by several months.  We now anticipate the public beta being ready by early 2009, with access to the private beta available towards the end of this year.

What’s Taking Us So Long?

For the most part, the delay is a natural consequence of our business strategy, which can be summed up as “Don’t quit your day job!”  Genlighten continues to be primarily a “nights and weekends” project for Justin and me and it will likely stay that way until revenue justifies otherwise.

Of course, we could put a lot of effort into getting angel investors to fund our early-stage development, allowing us to work full-time on Genlighten and hopefully get it launched and producing revenue a lot faster.  That’s a strategy we’re considering, but our team is well aware of the immense challenges and pitfalls involved, so for the time being we’re not ready to head down that road.

In the meantime, just in time for FGS, we do have a few updates to the Alpha site to announce.

Updated Flyers, Document Delivery Strategy, and Provider Criteria

First, the PDF flyers available on the site have been revised to reflect our new anticipated private and public beta availability dates.  The new flyers also reflect our intention to deliver most (if not all) documents as digital images via download from the site itself, rather than having our providers mail paper copies.  Accordingly, we’re looking for potential providers who have ready access to a digital scanner and are confident creating good-quality images from scanned documents.  These adjustments to our strategy reflect input we’ve received from several generous advisers in the genealogy research community.  Look for more details on this approach in a future blog post.

New Mailing Address

Second, we’ve changed our mailing address.  The new one is:

Genlighten
P.O. Box 893
Wilmette, Illinois  60091

This move doesn’t reflect anything particularly strategic.  Our previous commercial mailbox provider just happened to go out of business (about a month after we had business cards printed, unfortunately.)

Privacy and Security Policy

One of the main unfinished tasks from the initial alpha version of the site was developing a privacy and security policy.  I’ve noticed from our traffic logs that many people visit the blank placeholder page we’ve had up to this point.  I’m happy to be able to say that we now have posted our Privacy and Security Policy here.  Please take a look and let us know what you think.

Thanks for your patience…

I’m keenly aware that some of you have been following our progress for the better part of a year now.  I just want to let you know that we appreciate your patience and we’re working hard week-by-week to deliver what we’ve promised you.  Thanks for your ongoing encouragement and support — it means a great deal to our entire team.

I love it when someone at a genealogy conference stops by our exhibit booth and says “So, what exactly is Genlighten?” (usually while unwrapping a chocolate from our candy bowl). As you can imagine, I’ve given lots of different answers to that question over the past year.

How I answer the question “What does your site do?”

An example: I often tell a story about wanting to get an obituary for one of my ancestors from Jefferson County, New York and how nice it would be to find a local researcher in Theresa or Watertown who could track it down for me. “Genlighten helps you do that,” I say.

But most of the time I try to get out a simple, succinct “elevator pitch” — something like:

“We’re an Internet-enabled, human-powered search and retrieval network for genealogical documents;” or

“We connect you with local researchers who can help you find the genealogical records you’re looking for.”

Sometimes people get the concept right off, but often they don’t. They seem to need something to mentally compare us to… an existing business concept that they already grasp.

A “high concept” startup

Over time, I’ve tried to improve on our elevator pitches and craft a phrase positioning Genlighten as what Jeremy Liew of Lightspeed Venture Partners calls a “high concept startup“. That’s where you describe your business model using an analogy to an existing business that people already know well. The best I’ve come up with in this vein is probably:

“We’re kinda like eBay for genealogy document retrieval services.”

But there are several problems with that one. First, not everyone likes eBay, particularly lately. Also, Genlighten isn’t focused on bidding or auctions of genealogical services, so the analogy doesn’t really hold that well. Plus, at least one other genealogy-oriented startup is now using the eBay analogy — and it fits them better.

When I learned about Etsy, the online marketplace for handmade-crafts, I tried saying “We’re kinda like Etsy for genealogy lookups”, but few inside the genealogy community seemed to get the reference.

Tapping the wisdom of our exhibit booth visitors

On more than one occasion, visitors to the booth have come up with their own high-concept pitch for us. I heard it again a few nights ago at the IAJGS Conference here in Chicago:

“So, you’re kinda like Random Acts, only you’re not free.”

This one made me cringe the first time I heard it, at the FGS meeting in Ft. Wayne, Indiana over a year ago. It still does a little, though it’s actually starting to grow on me with time. It’s true, we are a little like Random Acts — we help you find people who can find genealogical records — and it’s also true that we’re not free. But we differ from Random Acts in several highly important ways, and those differences are part of why we feel justified charging for our services.

How is Genlighten different from Random Acts?

If you’re not already familiar with it, Random Acts of Genealogical Kindness (RAOGK) is a marvelous website that lists volunteers willing to perform genealogical lookups for free (or for just the cost of copies or gas). It embodies the spirit of volunteerism that powers much of the genealogical community: researchers help other researchers without expecting to be paid for their time.

I’m always honored to be compared to RAOGK, but of course our business model is significantly different than theirs! We want to help lookup providers get paid for their time and expertise (not just their expenses) in retrieving genealogy documents. And we aim to make money ourselves in return for the service our site provides. I discussed our “value proposition” in an earlier post. Here are some specific ways I think Genlighten will be different from RAOGK — different, that is, in a good way:

• We’ll provide a simple way to enable messaging back and forth between lookup clients and providers without the need to exchange e-mail or regular mail addresses. This should enhance privacy and security and help minimize spam.
• Each of our providers will have the chance to create a profile describing their genealogy background and experience so clients can make an informed choice when competing providers are available in a given area.
• When providers are out of town or on vacation, they’ll be able to temporarily put their lookup offerings on hold. That way, clients won’t have to wait for weeks wondering why they haven’t gotten a response to their lookup requests.
• The site will provide an online payment interface with state-of-the-art security, allowing clients to order lookups conveniently using credit cards or electronic checks.
• Providers will deliver the documents they find by uploading scanned digital images to our site. Clients can then download the the documents they ordered immediately without having to wait for them to come in the mail.
• Clients will be able to rate and review each of our providers based on their reliability, responsiveness, and customer service.
• The combination of client ratings, researcher profiles, and fees set by the providers will create a strong sense of accountability that will allow our users to order lookups from our providers with confidence.

Still in search of the right analogy

So we’re not really like eBay, and we differ in important ways from RAOGK, and we’re a little like Etsy but that probably doesn’t mean much to you. What then is our ideal high-concept elevator pitch? As you probably guessed right from the start of this post, I’m still working on it. And I’d welcome any suggestions that readers of this blog might have.

At genealogy conferences and in some of the responses we’ve received to our recent e-mails to genealogical societies, potential providers have occasionally referred to Genlighten as a “middle man” in the lookup process. Though the connotations of this term can be somewhat negative, it’s a label we’re willing to live with. After all:

1. We aim to make it easier for genealogical researchers to earn income by offering lookups.
2. We plan to receive a commission on each lookup request we process.
3. By creating an online lookup marketplace where genealogy researchers and lookup providers can find each other, we are indeed acting as a form of middle man (though we would probably prefer the term “third party intermediary”).

I’d like to use this post to clarify our intermediary role and explain the value we offer potential lookup providers.

Transaction Fees, but no Listing Fees or Subscriptions

We’ve decided to follow a “transaction fee” business model on our site, rather than a subscription model (such as Ancestry uses) or a listing fee model (which eBay uses). We chose that approach because it lowers the up-front risk to both providers and clients. A provider who offers lookups through us only pays to use our platform when a lookup transaction is successfully completed. If they don’t get any requests, they don’t pay us anything. Similarly, a potential customer pays us nothing unless they find a lookup offering they’re really interested in.

How It Works — A Numerical Example

Suppose the provider sets their price at $10 to retrieve a copy of a vital record from a local government archive. Our commission rate is 15% for independent providers (we discount it to 10% for societies). So when a customer orders a lookup and the provider successfully delivers the requested document, the provider would receive $8.50 and Genlighten would receive a $1.50 commission. [The customer would also pay a separate handling fee to cover the cost of payment processing.]

The Value Genlighten Offers

Why should providers pay us a commission, rather than keeping the entire $10 for themselves? Why shouldn’t they just “cut out the middle man”?

We think we offer a lot in return for the fees we receive. In particular, Genlighten aims to give our providers:

• A simple tool for creating both lookup listings and customized data input forms, so the provider gets specific, well-defined requests and just the right info they need to fulfill those requests.
• A secure online shopping cart that handles upfront payments via PayPal, credit cards and electronic checks.
• A robust customer feedback and rating capability.
• Targeted advertising on genealogy-relevant sites and searches so prospective clients can easily find providers’ lookup offerings.
• Aggregated world-wide demand so providers will have enough requests to make the regular trip to the local archive profitable.

Are these benefits enough to justify our fees? We sure hope so. And we’d love to hear what you think, one way or the other.

['Middle Man' image from Phil Hatchard's photostream on www.Flickr.com.]