The "Q1 2020: Splunk Ideas" blog is officially live! This blog post is the first in a quarterly series that aims to educate and deliver status updates on "Splunk Ideas." In this post, I will cover the history and goals of Splunk Ideas and supply some information about our initial success. Next quarter’s post will focus on the lifecycle of an Idea, with details on our internal process of reviewing, considering, and prioritizing your ideas.
Making Decisions and Running Out of Ideas
Back in 2015, having recently moved from Sales Engineering to Product Management, I had just celebrated my first GA (Generally Available) release (6.0) of a revamped Adding Data workflow. The decision to prioritize this work was easy – it was as broken as it was ubiquitous. Following this release, I was handed the “Search UI” feature-set, and I could not have been more excited — I got to work on THE main interface which our users interacted with their data! Unfortunately, the excitement soon turned to dread as I realized that it was going to be exceedingly difficult to decide on which improvements to make and what new features to build. What I needed was to make a data-driven decision, and I knew just the product!
What my dataset felt like 🙁
Unfortunately, the only data I could find was a set of over 2000 (!) “P4” Support Cases – enhancement requests that I was barely keeping up with, let alone mining for insights. That much data SHOULD be a PM’s dream, but since many of these cases were just conversations between a customer and a support engineer, analyzing them programmatically was impractical. These requests covered a diverse range, from highly complex (time zones are MUCH harder than you might think) to known bugs, to what I liked to call ‘beer-brainstorms’ (imagine two data geeks coming up with “how cool would it be” features over a few pints of beer). Most frustratingly, there was no reliable way to group them (if 50 customers had asked for the same thing). My solution? I used my prior experience helping hundreds of customers as a sales engineer to help me choose requests that I believed would have the greatest impact – better than nothing but not remotely ideal. After 3 years and many features delivered, I could feel myself running out of ideas that were born out of solving customer problems, so I decided to return to Sales Engineering. (There's a pattern emerging...)
Voice of the Customer
In 2019 I began to work with one of our Principal Architects, Jeff Champagne, who was expanding our Voice of the Customer program. In addition to improving the way Splunk managed relationships between our customers and our products organization, he was building a platform for supplying quantitative analysis of customer enhancement requests. Knowing (painfully) the immense value that such a platform would provide for both customers and product managers, I told him I wanted in! The platform was going to be named Splunk Ideas, and its two primary principals were:
1. To provide our Products organization with quantitative insights derived from customer feedback, to help make better decisions.
2. To provide our customers with prompt and realistic feedback regarding if, how, and most importantly when we would address their feedback, in the form of new and improved features and products.
To address the first principal (quantitative insights), we implemented a voting system so that our customers and community could collectively endorse a single idea. To assure that all our customers and community members’ ideas voices were represented fairly, we also devised a strategy for weighting votes based on their cohort and provide idea allowances on a per-feature/per-cohort basis. The details of this process will be shared in my next blog post.
As for the second principal (prompt and realistic feedback), we met with every Product Manager to determine what the ideal categorization schema would look like for their product & feature areas, as well as a commitment to consider a predefined number of ideas each quarter. We also implemented an internal monitoring & notification system to ensure that everyone is performing a thoughtful review and providing quality feedback within 3-6 weeks. (This timeframe is not a suggestion – it is a Service Level Objective).
Splunk Ideas Launch
In February of 2020, after a 6-month beta that generated over 300 ideas (Thank you, SplunkTrust!) we launched the general availability of Splunk Ideas to a very hungry audience. In just two months since the launch, we have over 600 ideas and more than 15,000 votes, from 100 distinct categories across 20 products. Our fantastic community managers review every idea and comment to ensure the submission guidelines are being followed, and everything is passed to a complex cohort-ranking calculation (in SPL, of course) that determines our “Top Ideas” monthly.
On April 1, 2020, we supplied our first set of Monthly Top Ideas to over 30 Product Managers. The internal reception has been overwhelmingly positive – it turns out Splunkers love making data-driven decisions! Over the coming weeks, our PMs will be reviewing your ideas and making decisions on if/when they can prioritize work to deliver solutions based on your needs. Over the coming months we will be expanding the scope of Splunk Ideas to include the collection of additional types of feedback, from early product development stages to beta releases.
Initial Success & Recognition
As of the date of this post, we have promoted 144 of your ideas across 11 products for review by Splunk Product Managers. I am even more excited to share that 13 of these ideas are already in development! (Many more to come, as our PMs are still evaluating their Ideas).
Distribution of ideas (Under Consideration) by Product
Launching Splunk Ideas was a collaborative effort and will only continue to be more so as we expand the program. I would like to wrap up this post by recognizing the outstanding contributions of a few notable teams and individuals, without whom we would not have been as successful. Specifically, I would like to recognize:
- Jeff Champagne for laying the foundation of Splunk Ideas and entrusting me with its reins.
- Charles Elliotte for managing and shepherding the multitude of simultaneous, cross functional tasks from inception, to beta & launch.
- Patrick Pablo, Evania Zhang, and Anam Siddique for their tireless reviews of all submitted ideas and comments
- Fred De Boer for developing our Splunk Connector for Aha!
Documentation & Links
I will be supplying more documentation in subsequent posts, but you can find a few resources below: