Dave Naffziger’s Blog

Amazon announced their first EC2 price reduction since EC2 was launched in August 2006.

Amazon’s price reductions take the form of EC2 ‘Reserved’ Instances. Users can opt to ‘reserve’ an instance for an up front cost. Reserved Instances then have a lower rate for each instance-hour use afterwards. For example, a Small instance normally costs $.10 /hour. A 1-yr Reserved Small Instance costs $325 up front and then $.03/hour. That works out to a 33% discount if you use it for a full year. A 3-yr Reserved Instance ($500 up front), results in a 50% discount over the 3 years.

The discounts are the same across the spectrum of EC2 Instance types. If you use your servers full-time, the 3-yr Reserved Instance is the better deal compared to the 1-yr instance if you use the instance for longer than 12 months:

What’s wrong with discounts?

Don’t get me wrong, the Reserved Instances will save BrandVerity a lot of money. I completely understand why these long-termish commitments make sense for Amazon, but some of the magic of EC2 just disappeared:

• Moore’s Law Violations: EC2 is nearing 3yrs old and Amazon hasn’t lowered prices (or increased capabilities) of its basic compute unit. They’ve introduced a ton of different configurations, but core prices haven’t budged. Committing to 3-yr usage and prices gives me the feeling that we aren’t on a pricing slope that parallels Moore’s Law.
• Oh, The Democracy!: Capital suddenly matters. Nothing costs more for the boot-strapped startup, but the company with capital can build a significantly lower cost-basis.
• Scale Flexibility: Committing to specific instance types increases the cost of moving between different EC2 types. UnReserved Surge capacity is now much more expensive vis-a-vis Reserved every-day servers.

Of course, Amazon isn’t bound by many of the assumptions I made above. Their next pricing update could address some of those issues - even for customers that have already purchased a reserved instance.

Our writeup and pictures from New Zealand are finally online (close to a month and half after we got back)! Two of my favorite pictures are below.

While there are tons of measures of a startup’s success, one of the more amusing ones is the degree to which your competitors take notice.

We have been fairly quiet in public about BrandVerity and our growth, so I always take a special interest in how new prospects find us (good practice at any stage of a company). We make trials incredibly easy to start, so we often know very little about the party on the other end that creates the account. Occasionally, someone will create an account that just doesn’t seem to fit with what we do.

So far we’ve had at least two accounts that I can directly trace back to a competitor. In one case, the CEO of a peer opened an account using an email address from her husband’s company. She never contacted us directly, but looking at her usage patterns in the site it was pretty clear that she was conducting a feature comparison.

The more recent one was particularly brazen though. An employee of a larger competitor created an account (using a hotmail address), and then sent a note asking to speak. The employee represented themselves as a marketing consultant that managed campaigns for several brands, but over the course of the conversation it became pretty clear that she wasn’t a typical customer. Questions about unrelated aspects of the service, size of company, and customer counts just didn’t fit with a typical customer’s behavior. I declined to answer the more sensitive questions, and after the call I created an excuse to call her back to clarify something. The response at the other end of the call was ‘Hello, ’.

Competitive research is an imperative in business and I certainly expect competitors to try and learn as much as they can about us. I actually think that more direct approaches yield better success (eg just reach out and introduce yourself), but as a startup you should just be aware that anything you say to customers and prospects becomes generally available information.

Oh, and it is great if your competitors accept your Terms of Service!

While speaking to a friend that is starting a company I realized that I haven’t written about filing an 83b Election. The election is a very important step in new company formation and if forgotten can have painful tax consequences.

I use stock vesting to refer to both traditional options vesting agreements and the buy-back restrictions on granted stock. In practice the two approaches generally serve the same purpose: to grant stock rights to an employee over time.

What is an IRS Section 83B Election?

The Section 83B election allows employees to change the tax treatment of a restricted stock grant. Normally, employees pay regular income tax when the stock vests (the restrictions lapse) and no tax when the restricted stock is granted.

The 83B election allows employees to pay income tax on the initial grant instead of paying tax when the stock vests. More specifically, they pay tax on the difference between the amount they paid and the Fair Market Value (FMV) of the stock.

Since the FMV of a brand new enterprise is generally close to zero, most founders purchase their stock at FMV. If the founders file an 83B, they end up paying no tax at purchase and no tax when the stock vests. They will only need to pay capital gains tax when the stock is sold.

The 83B election only applies to restricted stock – it only deals with the recognition of income on stock that has restrictions that lapse.

Let’s use an example to illustrate the effect of making (or not making) the election. We’ll consider a founder that is subject to a 4 year annual vesting schedule:

At formation, the company stock is determined to have a FMV of $0.001 per share and he is granted 100,000 shares. The firm grows and the FMV increases to $.10 in Y1, $1.00 in Y2, $10 in Y3, $100 in Y4 (nice growth curve). Assume a 40% regular income tax.

Initial Stock Purchase: $100
Value of stock that vests in Y1: $2,500
Value of stock that vests in Y2: $25,000
Value of stock that vests in Y3: $250,000
Value of stock that vests in Y4: $2,500,000

83b election filed
Taxes Due at Purchase: $0
Taxes Due at Vesting Intervals: $0
At the end of the 4 years, the founder owns all of his stock outright and has paid no taxes on it. Should he sell the stock, he would be subject to long-term capital gains taxes.

No election made
Taxes Due at Purchase: $0
Taxes Due at Vesting Intervals: Y1: $1000, Y2: $10,000, Y3: $100,000, Y4: $1,000,000
Total taxes paid: $1,111,000

And these taxes had to be paid before the company ever had a liquidating event (the founder never received cash for his stock).

Even worse, if the company collapses in Y5 the founder will have paid over $1M in taxes and never received any cash for his stock.

When do you need to make an 83B Election?

Within 30 days of assuming stock ownership. This rule is notoriously inflexible. The election form must be sent to the IRS within 30 days of the grant.

83B applied to startups

The 83B is typically relevant in two scenarios:

1. Restricted stock agreements for founders. The most common scenario is the restricted stock agreements of founders. Founders should usually draft and sign restricted stock agreements at the same time that they purchase their initial shares.

   A common scenario is one where founders don’t draft vesting agreements when they form the company. Some time later, the company takes investment and the investors require that the founders accept restrictions on their existing stock.
   Since the stock in question was granted more than 30-days ago, filing an 83b might not be an option.

   I’ve received distinctly opposing views from lawyers on this scenario, and a recent ruling (Revenue Ruling 2007-49) by the IRS provides some, but not necessarily full guidance.

   Creating stock vesting agreements after the 30-day 83B window is tricky and can require new stock grants that dilute the initial holdings. Altering existing vesting agreements is much more straightforward and is directly addressed by the IRS ruling mentioned above (and typically allows the original 83b filing to remain in effect).

   In my opinion, founders should always draft vesting agreements when the company is incorporated (for a multitude of reasons). This has an important benefit of giving them the opportunity to file their 83b election within the 30-day window. This significantly lowers the risk that they find themselves owing tax as shares vest if they are required to accept vesting agreements in the future.

2. Early option exercise plans. Some option plans allow employees to exercise their options prior to vesting. This provides the employees the tax advantage of starting the clock on long-term capital gains early. Once the options are exercised, they typically become subject to a stock restriction agreement and the employees would need to file an 83B after they exercise.

How to make an 83B election
The IRS doesn’t provide a form for the 83B election, but the process is straightforward. From IRS publication 525:

How to make the choice. You make the choice by filing a written statement with the Internal Revenue Service Center where you file your return. You must file this statement no later than 30 days after the date the property was transferred. A copy of the statement must be attached to your tax return for the year the property was transferred. You also must give a copy of this statement to the person for whom you performed the services and, if someone other than you received the property, to that person.

You must sign the statement and indicate on it that you are making the choice under section 83(b) of the Internal Revenue Code. The statement must contain all of the following information.

• Your name, address, and taxpayer identification number.
• A description of each property for which you are making the choice.
• The date or dates on which the property was transferred and the tax year for which you are making the choice.
• The nature of any restrictions on the property.
• The fair market value at the time of transfer (ignoring restrictions except those that will never lapse) of each property for which you are making the choice.
• Any amount that you paid for the property.
• A statement that you have provided copies to the appropriate persons.

The form we used to file our 83B elections at BrandVerity is here. Most lawyers will recommend that this notice be sent via certified mail.

Disclaimer: I am not a lawyer, tax accountant or otherwise qualified to dispense with legal and/or tax advice. You should always consult with a qualified professional before making decisions regarding Section 83b and/or your stock plans.

As BrandVerity has grown, I’ve sought out web-based applications to run various aspects of the back-office. Without fail, I’ve been consistently disappointed with quality of these seemingly-mature applications.

I applaud the focus that many of these applications have taken. There are stand-alone applications that that handle invoicing, payroll, accounting (of-course), transaction processing, etc.

All of the web-apps appear on the surface to be inspired by web2.0. They have the right colors, fonts, ajaxy goodness, etc. However, they fail on either basic usability or on core functionality.

Here are a few highlights from my recent testing:

FreshBooks - Invoicing
A service complete with APIs for access that even snail-mails invoices… but doesn’t email invoices.

SurePayroll - Payroll
They handle state payments for Unemployment Insurance, but not Workers Comp deductions. Broken form fields that don’t accept the most basic of information (business identifier)

PayQuick - Payroll
Ajax forms won’t submit. Complete lack of guidance on what values fields expect ($ vs. %).

Oh, and PayQuick won the PC Magazine editor’s choice award with SurePayroll a ‘close runner up’.

Authorize.Net - Transaction processing
Took a week to get a ‘test account’. Menu system fails spectacularly in Firefox (menu items simply don’t appear, but you don’t know they are missing).

And then there’s the accounting packages. If you’ve ever dealt with QuickBooks online (IE only), I’m sure you wish for that time back as much as I do. It typically takes 3x the clicks than it should to accomplish anything.

Maybe the downturn has an upside? All the developers working on Facebook Apps and Social Networks just might end up working for these companies (or starting their competitors).

Despite the fact that I despise fanboyishness, I’m developing a deep admiration for Amazon’s work with AWS. This admiration only deepens when I compare their ‘cloudy’ approach to Google’s.

App Engine reminds me of the Transformers that I used to play with as a kid. They were way cooler than any other plastic toys available, but they were still plastic toys. They had some flexibility, but ultimately you were playing with either plastic robots or machines.

Amazon’s Web Services remind me of the Erector Sets I adored. They came with electric motors, gears, pulleys and tons of structural pieces. I felt that you could build anything with an Erector Set. Not only that, but you could combine pieces from the set with just about anything else by bolting it together.

I’ve toyed around with both services and even built a simple tool on Google App Engine that we use for Regular Expression development at BrandVerity: PythonRegex. However, I increasingly find that App Engine isn’t an option for any project of reasonable scale or complexity (not to mention that I find the SDK cumbersome and unstable). It seems ideally suited for the next housingmaps.com, but not for most projects.

However, Amazon Web Services offers an incredibly deep and powerful set of tools that only continues to expand in scope and capability. Browsing the available AWS services feels a lot like a visit to Home Depot - the potential excites me. The building block nature of AWS allows the services to be mixed and matched with just about any application. EC2, S3, SimpleDB, Cloudfront, mechanical Turk, are all serving to drastically lower the cost of starting and operating a web business. Not only are the services super-useful but their very nature encourages a number of extremely beneficial programming practices (particularly dependency reduction and linear scalability).

I love that the pricing is incremental. My bill for toying around with AWS was $0.14 in the first month. I’m always wary of Google’s ‘free services’, and find that they require a careful read of the TOS to determine if it is a good idea to lock into their services. With AWS, I know my prices up front and can build very simple forecast models. Nothing is hidden. Not only that I love that AWS is constantly lowering prices. Prices have dropped for nearly all the major services in the last year.

Beyond the automated tasks, Mechanical Turk is purported to have 300,000 ‘Turkers’ earning extra income performing straightforward tasks that automation can’t solve well. Apparently the average lifespan of a task is less than a few hours. Photo tagging can be accomplished for 1 to 2c per photo. Semantic Web applications could be built with slight technological investment and instead depend on the semantic abilities of a small army of people.

The announcement that really impressed me was the aggregation of public data sets (Elastic Block Stores). I’ve long believed that the data sets maintained by governments and similar organizations possess a tremendous amount of knowledge that when freed from access constraints could expose deep knowledge. Amazon removed a significant barrier to this analysis and I can easily see a future where the world’s collected knowledge is available for easy analysis.

I eagerly await new developments from our new overlords in the cloud.

It was kind of cool to see Quova float across my VentureBeat feed late last week. Enterprise companies rarely cross the startup feeds, particularly those that are infrastructure companies. It is also amusing to see how the media reports news from companies you are actually quite familiar with.

The VentureBeat article is here and Quova’s press release is here. Other than making up a competitor that doesn’t exist (DoubleClick), there wasn’t any real analysis in the VB article or facts that couldn’t have been found in the press release.

The on-demand geolocation from Quova is significant because it drastically reduces the entry cost for world-class IP geolocation. We had experimented with a similar service nearly 6 years ago, but never really had the business structure to really sell or develop the service. The company that will most likely be affected is MaxMind, a tiny company that produces ok geolocation data. They’ve sold primarily to companies that were too small to afford Quova’s technology and have been successful amongst the solo-developer crowd.

As more and more of our data moves to the cloud, it frequently gets scattered across different sites. As much as I love data, aggregating and normalizing it is a pain. I have been an advocate of Yodlee for nearly a decade, and have become so accustomed to it that I can’t imagine how people manage their finances or frequent flyer accounts without it.

Microstock (stock photos that cost <$10 per photo) has become an increasingly large business and is rapidly tearing the $2B traditional stock photography business to shreds. There are probably close to a fifty thousand individual photographers contributing photos to sites like istockphoto, fotolia and dreamstime.

If you are a photographer, you’ll submit your images to multiple sites: the sites are non-exclusive, it costs nothing to submit a photo and it increases your chance of selling the photo (which can be sold multiple times). But tracking the performance of your photos is very challenging - you need to download data files from multiple sites, normalize them and then somehow link the photo-by-photo performance data from the different sites.

Rahul and Casey Zednick just launched LookStat, a must use application for serious microstock photographers. Not only do they aggregate the data from multiple sites, but they also do the very hard work of normalizing your data. That means that you can see how much a single photo is making across all sites.

I fully expect that LookStat becomes an essential application for photographers. Congrats on the launch!