Under the Covers of a Startup Software Product Launch

On the heels of our recent successful on-time launch of the next version of our Local Marketing Automation platform, I though I would go back and revisit what we have accomplished over the past few months and some of what lead up to this initiative.

From a macro perspective, it might actually be considered almost textbook SaaS (Software as a Service)development: Build and prototype quickly, get customers early, get feedback, drive to early revenue, continue iterating the product features and re-architect the technology in a just-in-time fashion.  Then, of course add in your typical startup constraints - budget limitations, resource constraints, just to spice things up!

Let me first take you back to the dog days of summer in August 08 when we launch what might now be called the beta release of our Local Marketing Automation platform (although we resisted the convention of actually calling it a beta) with our first set of customers.  The application showed immediate value, and our user base grew quickly.  However, by December of 2008, it became obvious that from an operational and scale perspective we had a few issues that might get worse given our current trajectory.  Additionally, aggregated customer support call data indicated a number of recurring themes, in addition to feedback from the field with potential clients showing some leading indicators of potential cracks that might become problematic when our growth track continued.  All of these diverse data points were indicating a need for a larger revolutionary overhaul over the current evolutionary development model.

January 2009: In addition to our standard ongoing iterative development process, we began brainstorming ideas for how the application could be redesigned to make a revolutionary leap forward.  We did this with a user-centric approach, starting at the user interface and worked backwards into more of the technical aspects of the design.  Even though our product falls under the B2B application space, our focus was to designing it for a B2C user from a design and usability perspective.  We wanted to make the application intuitive and usable, without the need for extensive training.  We first looked at other sites that were popular on the web that had related functions and worked well in the consumer world.  We then took these design concepts, mixed them with our domain expertise and applied them to our product.  This first produced a few hand drawn mocks, followed with some more functional screen flows put together in PowerPoint.

Next, we shopped these around to our all teams internally - sales, marketing with a deep focused on the groups that interacted with the app and our customers directly (inside sales, account mgmt, support).  They were our best window to our customers without the time or luxury for extensive market research.  Along with this, we presented the conceptual flows to our current client base.  From here, we moved forward by presenting the concepts to a select set of actual end-user customers identified by our account management team.  Looking back at my files there were at least a dozen major revisions through these prototyping exercises with many many more minor tweaks.

Now that we had a good idea of where we were going, the next step was to engage our creative team, who then gave the life to our wire-frames as they would appear in the real application.  This also served as an exercise in breaking down our interface into named visual components to support a common language when talking with our technical team.  This turned out to be a critical factor in ensuring we could develop the solution at breakneck speed.

So, to give some perspective - we did all of the above over a 2.5 month timeframe through Mid-March as a side project, while we still continuing to do iterative development to enhance the existing product (being extremely careful not to develop things that would likely be thrown away with the upcoming redesign).

In March we started working with our technical team on design to see what the new concepts would mean to the underlying architecture (data and business logic).  It indicated large changes - but for the better.  We would be able to throw out a lot of things and build new structures that would support our customers more effectively.  (In the end 80+% of our data structure and code base would be rebuilt).

In April we started the first of two formal development sprints with stabilization periods between each that would span a total of approximately 12 weeks.  In that periods of time, the product and development teams (with the support of the rest of the company) carried out a Herculean effort of rapid design and development - making constant adjustments as new issues arose.

About 4 weeks into the effort we locked down our release date to June 15.  This was critical to support ongoing sales efforts and work with clients to support their internal planning.  Due to the large number of dependencies and the scope of the change, about 6 weeks in, we had to start managing the overall effort with more of a ‘traditional software project management’ technique to incorporate all of the other pieces required for a release of this magnitude (customer transition & training efforts, outbound communication, cross company testing etc.)  However, at the core we continued to follow agile development principals, and the agile mind-set.

The cut-over was planned on a weekend, to give us the most time to react to any issues with the deployment.  We used almost all of the hours in that weekend to upgrade, convert customer data and validate to ensure a successful go-live on Monday morning.  On Monday June 15 the latest release of our Local Marketing Automation platform went live! … breath.

In retrospect, here is what I would consider to be our 5 key success factors for this redesign:

• Relentless focus on the customer, and letting form drive function
• Domain knowledge backed up with real world experience from the beta allowing us to design/build rapidly and make fast decisions
• Fanatical focus on prioritization to time-box the effort
• A passionate product and engineering team with a ‘lets getr done’ attitude
• An all hands on deck mentality across the company to support the product release where and when needed

Within days of launch, feedback has started rolling in from the our support team and through our customer transition/training webinars being held.  Here is a quote our support team received on day 2:

“I just wanted to say thank you for what you’ve done with the tool.  I had a list of things that I wanted to see changed, and this new version has covered almost all of them.  I really appreciate the changes you’ve made.  It is a lot more user friendly.”

Of course you can never sit on your laurels for very long as constant improvement is a must to stay competitive, but at the end of the day this is really what product development is all about. A satisfied, excited customer.