Since the early summer, my team has been building a new product based on our interactive video technology (play the video here). In addition to the Flash-based video application and content management console, we also had to implement a robust analytics system. After going through the build/buy decision, I decided that it was in our best interest to “lease” Omniture Site Catalyst instead of building it ourselves (we also didn’t have the bandwidth to implement an open-source system).
Omniture is no doubt one of the most powerful analytics systems I have used, and it is definitely going to provide a lot of value for our customers. It is also going to play a major role in our customer development process internally. Now that Adobe has purchased Omniture, I assume that their strategy will be to leverage the power of the platform for measuring web based activity and conversions in applications using Flash and other Adobe technologies. Assuming this is the case, I believe Adobe has to minimally focus on the following issues if they are going to be successful:
1. Make it easier to be a customer
Disclaimer: I have to say that the Omnitureteam is one of the most professional, patient, and accommodating teams I have ever worked with. OK, on to the dirt…
I worked on being a customer of Omniture for over 4-5 months. I had a great sales team working with me, but it seemed that their hands were tied. There were two major blockers to becoming a customer: 1. No trial accounts, 2. No platform/OEMcontract.
I believe that Adobe would be foolish to continue the practice of “no trial accounts” at Omniture. Now that they have a public developer sandbox, this may not be a problem anymore, but it was painful trying to do a proof-of-concept using Omniture. I essentially had to become a Genesis Partnerso I could have early access to the sandbox. Luckily I had a good sales rep I was working with who knew how to work the internal “politics”, but it took much longer than I would have expected.
After doing a proof-of-concept of Site Catalyst, we began the contract discussions. The legal docs pretty much had to be rewritten in most areas because they really don’t effectively service customers who want to use them as an OEMor “cloud based” analytics package. Considering that a lot of Adobe’s customers are start-ups, they better make sure that they don’t require them to spend thousands of dollars in legal fees to redo a contract. I have to think that Adobe will fix this, and I believe they should make it as easy as becoming a Google Analytics customer.
2. Foster a vibrant developer community
I’ve been really impressed with the way Adobe has managed the community with their Open Source Media Framework, and I think they should apply the same level of effort to the developer community who are using Omniture as a back-end. While Omniture has forums, a knowledge base, and a fair amount of documentation for developers, I do not think it is easy for teams like ours to figure out the best way to leverage the power of the Omniture platform. There aren’t a lot of blogs out there describing best practices and the platform team seems to be significantly understaffed. I was able to make a bunch of calls to get some help from Omnitureconsultants, but that will not scale to the Adobe community. If they have every small start-up calling through the org chart to get answers, it will kill them operationally.
3. Re-architect and simplify for rich internet applications
Like most web analytics systems, Omniture measures everything in either page views or conversion events for funnel analysis. Sure, they’ve added support for video, but it isn’t tightly integrated to the rest of their system. With Flash, there are no page views and conversion events tend to be more fine-grained than traditional web workflows. For example, you might want to measure a conversion event when a person clicks an interactive overlay on a video. Omnitureis more set up for conversion through the classic shopping cart purchase process. Sure, it is infinitely flexible and can be configured, but it starts to get a little “hacked” after a while.
Another critical architectural change is making the data available in real-time. I realize that Omniture is more of a traditional “trend analysis” tool than a real-time tracking tool, but in order to compete in the real-time web, they are going to have to process the data faster than they can currently do (which isn’t terribly slow, but isn’t real-time). I have a suspicion that this would be impossible without a major re-architecture, so I’ll bet we won’t see this for a while. If they processed data in real-time, then they would definitely have to open up their API’sand not limit them by tokens or other API metering models. If you are a proficient Flash shop, then you’ll probably want your report visualizations in Flash as well. Therefore, you’ll likely need a robust, unlimited, real-time API for the best experience.
If Adobe is going to win the game as the preeminent toolkit and platform for the the rich interactive web, they definitely need a solid analytics offering, which I believe is the major reason they purchased Omniture. However, even with Omniture’s market leadership, they have a ton of work to do to make Omniture work seamlessly as this platform. If they don’t make these changes, I believe that Google Analytics will continue their innovations with event tracking, opening up their API, and eventually becoming real time. If they don’t get there in time, then start-ups in the vein of Mixpanel or KISSMetrics will fulfill the underserved market of start-ups who don’t have the resources to build their own analytics system.