From Failure to Launch
The Problem
While advising a privately held Internet company, a major platform relaunch project had gone sideways. The company outsourced the development of the new platform to a large NYC-based interactive agency. The project was originally budgeted at well over $500K with a potential buffer of 15%. The firm estimated a late summer launch, with a total project timeline of 8 months. 5 months into the project the agency announced they would need all 15% of the buffer and tacked on another month to the timeline. By mid August, close to the original launch date, the agency announced they required another $300K to finish the project.
This was the final straw between these two parties. The agency’s client, whom I was advising, called me and asked if I could step in and help launch the new platform. There was no way they were going to agree to any more cost overruns with the agency and no longer believed their estimates. I agreed to take on the project.
Morale was down throughout the company as the entire staff was greatly looking forward to launch of the new platform. Furthermore, many projects had been put on hold during the platform development, only adding to frustration when the release dates kept slipping. Therefore, I knew I needed to show the executives and staff progress quickly, get live quickly, and keep the costs to an absolute minimum.
The Solution
I applied a 3-tiered approach to solve the problems mentioned above. The three tiers were:
Lean and MVP
Upon getting the product handover from the interactive agency, I immediately did an analysis of all the incomplete features still in the development roadmap. I identified the features that I believed were non-critical and advocated for their removal from pre-launch development. This followed the lean concept of launching a minimum viable product (MVP) and not worrying about every bell and whistle one could dream up. After a few meetings with the company’s executive team, they agreed to drop what I estimated as 3 months of work from my new team and agree to launch a much leaner MVP.
Offshore
I have extensive experience leveraging offshore resources to execute projects and I used my contacts in Eastern Europe to assemble a qualified offshore development team in a matter of weeks. The team was talented and nimble. It consisted of a Senior Engineer, a part time System Administrator and a Junior QA Analyst. I led the team and contributed a bit to all three of these areas as needed. We had virtual standups on skype each morning and worked off the same agile board I set up in Jira. It took 3 weeks for before the team was at full productivity, but once they were our velocity was consistent and predictable.
Agile
Within Jira, I set up weekly sprints to make sure we were deploying changes regularly. This helped the project on many levels. It kept the offshore team focused on smaller scopes of work each week, helping us stay on target. It also helped me gain confidence in the longer launch timeline, as I was able to predict velocity with relative accuracy and communicate timelines with confidence. It also kept the business motivated as they saw updates to the staging environment weekly, improving morale and reigniting excitement in the project.
As the company started seeing more and more of the new platform on the staging environment we set up, they naturally had a number of change requests. Some of these changes were pretty fundamental to the platform and did cause an additional month of delay to the launch. However,since we were following Agile, we were able to incorporate these changes seamlessly into our weekly workflow.
The Result
Two and half months after we took over the project from the interactive agency, we launched the platform and sent 10% of the company’s traffic to the new site. After identifying and fixing a few critical bugs that were not found before launch, 2 weeks later we began to drive 100% of the traffic to the new platform. Despite losing 1 month to new change requests, we successfully launched the new platform 3 months after taking it over at a total cost of $80K, 73% less than the interactive agency’s last estimate. There were no major hiccups or outages during the launch phase. In fact, the marketing director for the company, who had been working with the Internet since the early 90s, repeatedly said he had never been involved with a smoother platform launch in his professional life.