Interview : OpenCandy CEO Darrius Thompson
OpenCandy (www.opencandy.com) is a startup from San Diego, California that allows software publishers to connect and recommend other pieces of software during the installation of their product. It allows independent software makers to create distribution revenue while maintaining a good relationship with their users or simply to recommend other pieces of free software that they like.
How did you guys come up with the idea for OpenCandy?
We were looking for a problem to solve that we would be passionate about, that would leverage our past successes, and that had a good probability of getting us to self sustainment in a reasonable amount of time.
Before, OpenCandy members of our team developed both Open Source and commercial software. We knew that it was challenging for consumers to find the *best* software for solving their particular problem and we knew that developers found it a challenge getting their products discovered. We felt strongly about solving this problem and that included finding new or additional methods for funding software projects or organizations.
For many developers this means giving them the potential, if desired, to go from a part time to a full time opportunity. The idea of being able to help an Open Source project be truly competitive within the commercial marketplace is exciting. The competition it creates is good for everyone. Mozilla’s Firefox is a great example of that.
Our background, which included all of the DivX software products, went from creating software itself to figuring out how to get mass distribution that could be successfully monetized.
We’re passionate about helping developers manifest the vision they have for their software.
How does OpenCandy work?
We have created a network of software publishers that offer great products. When developers are part of our network they have the ability to recommend other software directly from their own installer. Their software can also be recommended as part of the installation process of all the other great pieces of software in our network. All of this happens dynamically at installation time thanks to a small plugin we provide for the major installation platforms such as NSIS, Inno, etc.
How does the sign-up procedure work?
If a software developer wanted to recommend other software products it goes kind of like this:
1. They tell us they’d like to try out our service by coming to our site and filling out a quick form
2. We quickly review their product to make sure it meets our software guidelines
3. We create an account for them and provide a very small plugin that they can integrate into their installer (our Publisher Kit is also available for any developer to download online and play with without having an account)
4. They pick what software they would like to recommend. This includes products that may or may not offer a financial incentive. That’s up to them, and if we don’t have a product in our network that suits them, they can let us know and we’ll try and get it added. This is important since they know what their consumers may like much more than we will
5. We test it to make sure it’s working then we coordinate the launch with them
As part of this we also provide a few basic analytics tools that helps projects understand more about and improve their product distribution. More specifically it helps them improve their download and install experience for their users. They also get uninstall data which gives them an insight into how users are feeling about their product. If they are recommending a product that offers a financial incentive they will also get to see their potential income for the month.
If they want to have their product recommended it’s even easier.
They just let us know they’d like to participate through our form, we make sure their software meets our guidelines and we make their software available for recommendation. We’ll provide them an analytics console that provides them distribution intelligence similar to what you might see in a web analytics console.
The way recommendations work is when a consumer installs their software we decide which of the other software they have decided to recommend are valid (not already installed, support the right OS, etc…) and from there which one the user is more likely to appreciate. We then show them a single recommendation screen as part of the installer and if they accept it, we help them to download and install the recommendation automatically.
Have you guys come across any challenges since there has been questionable use of installers in bundling software in the past?
I think in most cases we’ve been lucky and thankful to get past this challenge as quickly as we have. We’ve been working with developers and organizations who really care about their product and users. The goal of our network is first and foremost about helping developers and consumers, and coming from that direction makes a huge difference. I think our partners can sense that we truly care about their products and consumers. This has helped us minimize some of the challenges here. Our first partners came from the Open Source Software community which provided early scrutiny that helped shape what we do and how we do it.
We also have a set of software guidelines that were influenced by StopBadware and Googles Software Principles. We run each publisher through a process that ensures they meet these guidelines before we launch their software into our network.
As OpenCandy grows, we’ll continue coming across challenges and hopefully we can address them in a way that finds the right balance between keeping consumers happy and helping developers get distribution and/or make money.
What have you done to address user privacy?
Privacy comes up most as part of the recommendation process and our analytics service. When we make a recommendation we want to ensure that it’s good. At a basic level we validate for the basics such as OS, Language Setting, Product Already Installed, etc… We can also validate more specific characteristics. For example if there is a great Outlook plugin they’d like to recommend we can first validate the consumer is an Outlook user. We can validate for characteristics that are specific to the recommendations being made. This means there is client/server communication occuring. We do this in such a way that the data transmitted is not personally identifiable and we only send back a yes/no response that is associated with this non personally identifiable ID (much like a web cookie).
We also provide analytics to our developers which helps them understand if they are providing an optimal and efficient download and install experience. We provide information that let’s them know when their product was un-installed. This can help our developers better identify when and why users are cancelling installations, not completing them, or uninstalling their products, and as we improve the experience together we can immediately see the results and continue to try new things. This is also done in an anonymous fashion. It’s, again, much like the web world where every site is running something like Google analytics and looking at data such as visitors, drop off rates, etc… but we do it in a truly anonymous way.
How long have you guys been doing this?
We started in February of last year and launched in October of the same year. We launched privately with a handful of select partners. It was important for us to launch with partners who had good products and were very brand conscious. We also knew that there was a ton of knowledge we needed to gain to help us ensure we shaped the product the right way and could provide the level of service that we wanted. We’re now at a point where we can start opening up our network to a wider group of users as we’ve been very selective in our partnerships thus far.
Are there other tools you provide software developers?
I’d classify the following as a sort of tool… What we do with each new partner is evaluate their entire process, not just items relating to OpenCandy, and work with them to optimize all of this. This includes improving messaging, their downloads, installs, registrations, sales, auto update process, etc… It’s really all about the user experience. Making it easy for users to understand what they’re offering and making it easy for them to try the product out. It’s easy to make this a hard process for users and it is hard to make it easy… but doing this can drastically change the goodness and size of a projects active user base. We’re helping to do this rather than just enable recommendations. This will helps us identify how we can better improve or add to our own product and services.
What else are you guys doing with the OSS community?
We recently sponsored and participated in the Libre Graphics Meeting where creators of graphic applications such as Gimp, Inkscape, Scribus, Blender, etc recently got together. Dr. Apps, our community leader, hosted a conversation at LGM entitled “Open Source and Money: Not Mutually Exclusive” although it got retitled to “OpenSource and Rainbows” right before the meeting! You’ll also see us at events such as OScon this year, discussing ways to help your projects get noticed and funded.
Beyond that we are directly involved with the community through our partners and potential partners we reach out to. We reached out to the OSS community before any other community, knowing that we’d get important critical feedback. Our vision is about helping developers create a better world for software and consumers. OSS is a big piece of this so we’re very active in creating 1:1 deep relationships with our partners.
I almost forgot to mention that in the final days of the Sourceforge Community Award’s nominations we had several projects using OpenCandy as a tool during their install process to get nominations for their software.
How does software moving to the “web” impact what OpenCandy is doing?
It becomes more important. We know that building software applications is becoming easier and cheaper which means there will just be more applications. The problem of recommendations gets much larger. We’ll see more developers struggling to find the right audience and we’ll see consumers lost in a sea of applications. We see this happening already and in a specific case with iPhone apps. It’s a great thing that we’ll have more choice but as we see this growth we want to be the solution that helps consumers find the right applications for solving their specific problem.
Certainly specific types of software belong on the desktop and other types of software will just take longer to transition. Either the technology doesn’t exist yet to make the experience good enough for the user or prevents feature parity. The economics also have to work. We can’t use Final Cut Pro on the web when dealing with very processor intensive functionality and massive files and a requirement for quick application response.
Even today we are working with partners here who have web services but also have applications. Great examples are companies such as online storage providers where they have a downloadable application and that application is attached to a web service.
What do you guys have planned for the future?
Our vision is to assist developers by helping them find the right users for their products. We’re looking at creating more active ways for consumers to participate and engage in this. The recommendation engine that we’ve built for our network was designed with this in mind. It can utilize active consumer feedback to improve recommendations and manifest them in other places beyond the installer screen. Our next step is to increase our consumer touchpoint in a way that actively helps consumers discover great products. Beyond that we’re really focused now on growing and shaping our network and with growth we’ll experience new challenges for sure.
Something I’d love to see happen is helping a great product with minimal visibility get some serious publicity through our network. If you look at some really great applications out there which are getting millions or even hundreds of thousands of downloads a year and imagine if they used just a small part of that to recommend a great application which wasn’t well known in the community. There are a number of great applications getting a few hundred or thousand downloads a month and the impact here could be huge… how about helping find and create the next Firefox.