Inspiration from the Microsoft Accelerator for Windows Azure, Powered by TechStars

A trip to the Microsoft Accelerator for Windows Azure last week reinforced my belief that cloud computing has made this simultaneously the most exciting and the most competitive time to be a startup.  The consumption-based cloud model lets you quickly spin up and experiment with your ideas and then if it takes off, great, pay as you scale.  If it doesn’t, okay, just dial it down.  Cloud computing makes it an exciting time to be a startup because so many people can achieve the level of access and capacity that they need to innovate, build, pivot, and launch.  Cloud computing also makes it a challenging time to be a startup because that lower barrier to entry means more people can play at the party.

Even with the increased competition, I really envy the ten companies in the Accelerator and the thousands of others in similar stages that are leveraging Windows Azure to solve a wide variety of problems.   These ideas range from how you can pay bills faster from your phone to crowd investing in the real estate market to behavioral analytics for email.  Even though there is a lot of variety in the class, there are some common challenges to each of the companies.  After meeting with each of these start-ups, I thought I’d share some guidance that (I hope) is relevant not just to these ten companies, but to most tech startups.

Build Cloud First

The first thing I want to call out is the huge advantage there is to starting clean and building cloud-first with PaaS.  Having talked with hundreds of early phase startups, there is a very clear pattern: startups love platform-as-a-service.  The reasons are simple: it’s faster, it’s simpler, it removes a lot of stuff they don’t want or need to deal with (like patching an OS), and it provides abstractions that allow startups to not only move quickly but to leverage capabilities that they may not be expert in (such as media streaming or pushing notifications to mobile devices).  When I was writing software for similar phase startups years ago, I would have *loved* to have had these types of higher level building blocks – particularly for an amazingly mediocre programmer like me!  PaaS is a startup’s best way to experiment, live lean, and work quickly through build-measure-learn, to quote the Lean Startup. 


The winner of our ‘Fast’ contest, Magnus Martensson, can tell you just how valuable building cloud-first can be.  After all, it saved the guy’s marriage!

Fast Fail with Your Business Model

Along these same lines, it’s crucial for early stage startups to keep in mind that it’s just as important to fast fail with business models as it is with software.  For developer-focused offerings—the audience that a few of these startups are targeting—there are three to four very basic business models (with a wide range of creative variations and possibilities). The most popular three approaches I have seen are the ‘freemium’ model, the licensing per <seat/core/server/etc> model, and the subscription model.  Finding the right fit for your business is going to take some trial and error, but that’s not a reason to be discouraged.  You probably won’t completely nail the product in v1 so there’s no reason to expect you’ll nail the business model in v1 either. 

MetricsHub provides some advanced monitoring, metrics, and scaling options for other applications running on Cloud Services.  They’ve tried a few business models but currently offer pricing tiers that correspond to the number of instances they’re monitoring.       

When you’re considering the pros and cons of these different options, there are a few things to keep in mind.  First, even though the freemium model is popular right now, the market doesn’t always equate free with quality or value—the assumption can easily become that you don’t know what you have yet so you’re giving it away.  To counteract that perception, you need to clearly articulate the value of what you’re giving away, and be clear and intentional with customers about when and why you will ask them to pay.  Second, if you are targeting the enterprise, when you use a licensing model, you might end up needing a pretty substantial sales team in order to scale.  If the business model growth is tightly coupled to salesperson growth, you should think hard about alternative or complementary direct sales models – or wait until you hit a certain user or revenue growth target.  Third, subscription models naturally become tiered.  When you have different tiers, it can be a challenge to move your customers up through tiers if you haven’t made the value prop distinct in each level.  Just think: when was the last time you called Comcast to adjust your channel package?  That makes on-boarding customers in the appropriate tier mission critical to your business intentions. 

Go to Market (GTM)

Product and business models are two places where you want to fast fail.  The highest level questions you need to answer around go to market are as critical as the software you are writing – you should be thinking about your product and GTM in one frame of mind.  At the most basic level, you need to know whether you’re asking your customers to make a significant change from their current behavior or whether you’re asking your customers to let you improve how they’re currently operating.  You can be successful with either approach, but that’s something you need to figure out early because it’s the foundation of the core value prop – and therefore the ‘heart’ of the product and marketing. 

All ten members of the Accelerator class have really done a good job in achieving some clarity here.  That doesn’t mean they all fall into one camp or the other. Realty Mogul is really clear that they’re aiming to drive a behavior change.  That’s what positions them so well to land an investment diversification message — particularly, I think, to those who would otherwise exclusively make angel investments in early stage tech startups.  


Mobilligy, on the other hand, lands squarely in the “improving customers’ current lifestyle” camp.  That focus allows them to land the message that their customers shouldn’t change their lifestyles based on when bills are due, but rather to change their bill payment schedule to fit their lifestyles.  The important thing is to make that decision early so that your messaging doesn’t get muddled.  You have to identify your most basic ask so your customers know what to say “yes” to.  After that, you can go right back to the fast fail model on specific, lower level messaging.  In other words, if you find yourself A/B testing your core value prop, you might need to take a step back and reset the compass.    

Customer Evidence, Customer Evidence, Customer Evidence

The most valuable words aren’t your own.  It’s true for new companies; it’s true for established companies.  Customer voices are the best ones to tell a story.  For new businesses, however, there will be fewer customers to call upon and point to.  Don’t let that stop you from taking every opportunity to evangelize your customer wins and how they’re using your product.  It is impossible for you to use too much customer evidence and it is foolish to regulate your use because you have to point to the same customers over and over.  Although you may be tired of hearing the same story, it might take years to get that customer story broadly out to the world (unless you have big dollars for marketing or the ability to have Bieber tweet about your amazing new widget – right?).  Rinse and repeat, get the most illustrative and exciting customer stories down and tell them again and again.        

Closing Thoughts

Experiences like this remind me why I love being able to work on the platform that is helping power these startups and hundreds of thousands of others around the world.  I’m also thankful for the opportunity to see firsthand how much of an impact programs like BizSpark and TechStars can have.  Both the BizSpark and TechStars team have really made it their top priority to connect these teams to the people and resources they need to succeed.

As you can probably tell by how fired up I am in this post, I think I picked up some of the incredible energy from the Westlake offices where this Accelerator class is based.  And a great local benefit to our Seattle Azure Accelerator?  Right across the street is some of the best pizza in the northwest at Serious Pie.  Great minds building exciting new software – with some really good ‘za. Yep, it’s a great time to be a startup on Azure.

If you’re currently working at a startup or have a great idea that you’d like to build into a company, I encourage you to apply for our next Accelerator class here.  We’re accepting applications until February 1, 2013.


Bill Hilf
General Manager
Windows Azure