Written by GoodData Author |
As the countdown clock ticks closer to embedded analytics products’ launch time, there are specific tasks that every product manager must complete. You may already be familiar with the multi-step project management pre-launch SaaS product checklists that have been making the rounds for years. You might have even downloaded a few to help you manage your embedded analytics project’s details.
We’re not going to go through every point on those lists. You’ve likely already checked off most of them via your enterprise data product or business intelligence product’s release process. Here we’re going to highlight a few that are important and sometimes missed when launching a data product.
Arrange an Independent Review
You and your team members are too close to this project. You know how everything is supposed to work and that means that you are more likely to gloss over bugs and and errors--or not notice them at all! Before the big launch, have an independent reviewer (or your red team) look over the product with fresh eyes.
Conduct Expert Testing
Depending on the point in your development process you may not have the time or the budget to hire someone completely new to do this. Your best option at this point, according to James L Hutt, is to find someone who already has experience building data products, like another product manager, to give your product a final test. Bonus points are awarded if you can find someone within your company to do this.
Why? Because this person will likely have worked with the members of your team on other business intelligence projects before. She or he will know where to hunt for bugs based on past performances and the test should go much quicker than if you were to hire someone to test “from scratch.”
Set Up Usage Tracking
Sales numbers are just one of the data points that you’ll tracking post-launch, especially when you trade in an old data product. You’ll also want to track your OEM analytics to see how each component of your SaaS performs for buyers. To do this, you’ll want to test your embedded analytics tracking prior to the product’s big launch, to make sure that all of the code plays nicely together.
You should also define what quantitative success looks like for your data product. Start by setting a timeframe: 3, 6, 12 months. Next, determine topline engagement and adoption KPIs: Accounts Accessed, Daily Active Users, NPS, Report Executions, Returning Users, Session Length. After high level metrics are set, dive into the details: Which dashboards will see the most traction? What adoption levels will warrant where you invest in the next release? Which metrics do you need to track percent change? The more granular the better, as hitting topline KPIs and missing low level metrics can give you valuable insight and dispel assumptions that will guide the direction for phase 2 of your data product.
Test Your Dashboards
You need a central space in which to compile and analyze all of your launch (and post launch) usage data. The best way to do this is to create a series of executive dashboards through which you can scan quickly for data, problems, feedback, etc. Do a test run of each of your dashboards before you officially launch your product.
Quadruple Check The Website
By now you’re probably a little bit sick of your data product’s website, but it’s worth taking another look through everything before it goes live. There’s always one last typo to catch. Smashing Magazine put together a great list of the details you should check during this last read-through and test.
The Soft Launch
After you’ve beta tested, A/B tested, and done a final checkup via a fellow business intelligence project manager, it’s time to do a short soft launch of the data product. This is where you allow public access to your project, either by rolling out your analytics feature to users in waves, or by “turning on” the feature without a formal announcement. Either way, this allows you to test out all of your internal processes with external servers and users. You can offer the product (for free or at a discounted price) for a limited amount of time--usually just a few days ahead of the Big Launch--with the caveat that the software may still have a few kinks and encourage users to send you detailed feedback about your features’ performance.
Implement Soft Launch Feedback
Distribute the initial feedback you’ve gotten to your team so that it can be implemented to ensure that your BI is ready for the Big Launch that’s coming up.
The Night/Day Before
The night (or day, depending on your product’s launch time) before you officially launch the product, run through most of these steps again. Do a final test of the product, scouring the code for bugs. Test upload/download time with your cloud servers. Make sure your dashboards and your executive dashboard are running smoothly and tracking all of the pertinent data. Fix any bugs or glitches.
Freak Out a Little Bit
This step is self explanatory.
Push the big red button.
Congratulations! You’ve officially launched your latest BI product. Now you can take a moment to get a sip of water (or something stronger) and then it’s back to your desk to start tracking and analyzing the embedded analytics and OEM analytics data that will be coming in from your buyers.
Written by GoodData Author |