FTD had a systems-based problem and opportunity space ripe for disruption. Production of assets is costly for any organization. Hiring a photographer/videographer, talent, booking a location, set design... And the output is generally a single purpose solution. Advances in computer generated imagery (CGI) have enabled the production of assets that are indistinguishable from assets produced in studios as seen in Hollywood. However, the cost of the tools and the expertise needed to use them remained out of the "consumer level" domain. Until now.
I built a business unit to "model" assets for generative reusability. We started with strawberries to prove the concept and to address the need for customization in ordering sets of berries on berries.com. There was a lot of skepticism about the viability of using CGI and if they would look "fake" vs images made in the studio and post-processed.
We needed to educate the organization about these new capabilities and clearly show how we would automate the production of assets that were unique and "real". Using procedural techniques, we designed a system that would allow FTD to control all characteristics of an asset, including imperfections.
Anneka Cady and Lindsey (Steimle) Wekenman are experts in CGI production. Our team presented to the CEO and down to show how we would create a system that would enable a myriad of services and media assets. The artists provided an "under the hood" view of how easily procedural techniques (generative media) could be enabled.
To address speculation, bias and skepticism, we designed a "visual Turing Test". First we conducted an "image lab" where we showed side-by-side an image done using CGI and an image produced in a studio. Image lab results were impressive, with internal and external audiences identifying the studio-produced image as "CGI" and the CGI produced image as "studio produced". We then launched an A/B test on the site during the busiest time of year (Valentine's Day). This test validated the results from the image lab and produced higher conversion as well.
Reporting to the COO and CEO, we needed to operationalize and produce to scale quickly. We proved to be able to reduce the cost per media asset by 80-90% (including video, lifestyle, etc).
The solution addressed more than the production of media assets for the core web properties. Local florists were bound to prescriptive bouquet designs from a print catalog distributed by corporate. Local florists were complaining about this because the joy of floral design is customization and creative flair. Another issue acute to the organization was returns or customer complaints that what they saw and ordered online did not match what they received. This caused additional inefficiencies and cost to the organization. Local florists need to move inventory that has a shelf life of 3 days to 1 week. Thus, they lose over 50% of their inventory monthly. Unlocking the ability to create assets for their websites based on inventory awareness would provide exponential value to both the organization and the local florist.
Working with Anneka, the team produced a video using AfterEffects to produce a "sizzle reel".
Additional opportunities presented themselves, like the ability to create revenue streams by "renting" or selling the rights to various 3D models to consumers and production studios (gaming, movies...).