The Decision To Open Source ArrayFire

July 31, 2014

in Engineering, Marketing, Pricing, Product Management, Sales

We have decided to open source ArrayFire, our primary commercial product. We are confident this decision will result in better outcomes for our customers, free users, and our entire market.

Our current market is rife with mediocre GPU and accelerator software tools that merely get airtime because they are open sourced. ArrayFire is hands down better than all the alternatives:  faster, more features, greater reliability and robustness, nicely scalable, and truly flexible in choice of hardware target. For the past 7 years we have been working on commercial proprietary software. That single fact alone has greatly limited our ability to serve our market.

Let me take a minute to address 3 natural questions that have weighed on our minds and hearts as we have come to this decision. I’m phrasing these questions as though a business adviser is grilling us about our decision:

  1. Who cares about providing software to users without money? They will never pay you anyway, so what do you care?
    1. It is all about the network effects as I wrote yesterday. Discounting the value of non-paying users is a terrible mistake. There are so many ways to monetize open source today, especially in the software tools space. Free users are monetizable indirectly as I’ve written before. We will not cling to dying business models that rely on proprietary software license sales.
  2. Maybe you just need to improve your sales efforts to reach your potential.
    1. Our sales efforts are wonderful! We have had a couple of products in the history of our company. We have applied the same sales techniques to each of those products. They have produced great results on previous products and are proven. In Atlanta, we are surrounded by the best sales and marketing companies in the world. We kick butt at sales and marketing in our space relative to all our competitors. Yet our direct product sales have not been growing as fast as we would like. After some deep introspection, we conclude that internal changes are not going to move the needle. Rather, our shift to open source business models will make a huge difference in our ability to grow as the market gets to use much better software with an easily adoptable open source software license. So we are going to shift full throttle to the future.
  3. Won’t potential acquirers have less appetite if you go open source?
    1. We have acquisition talks every year with various partners. Until they have teeth (e.g. a term sheet), they are not relevant. Potential acquirers have to live in the same market that we do and need to make the same decisions for growth and success that we are making. We believe that by building a solid and growing business, we are doing the right thing for everyone in our market.

We have debated keeping part of the current ArrayFire proprietary. We have debated limiting use for some people. We have debated the type of software license. In the end of all these debates, we come back to the truth that we want to be as open and permissible as possible. We are open sourcing everything. No strings or limitations of any sort.

Ever since we made this decision, we have been felt very confident about our growth plans. We are excited to share our baby with the world.

Tomorrow I’ll share thoughts on how to get from closed source to open source.

What are your thoughts on open sourcing your software? Would it help your business?


Previous post:

Next post: