Phoenix from the ashes: Re-crafting a legacy system to design a new one
Note: The full case study, looking at the design decisions involved in creating Adomatic.io, is currently in the works. In the meantime, posted here are some of the results of those decisions made along the way.
What led us to build this product?
In an effort to increase revenue through our legacy platform (Oomph) and to diversify from the shifty publishing market we began selling the process of manually converting thousands of print ads to digital for use on mobile devices. By many measures it was a success, we gained new revenue sources and gained a whole lot of knowledge about how to effectively transfer content from medium to another without an approval process.
Because we spent so much energy in documenting the design decisions, it was only natural for us to want to transform those decisions into an automated process allowing it to scale, serving many more people than we could ever do with a team of designers.
Despite the confidence of the big print publishers, they turned out to be a disappointment not only because of their decline in rich media production but also because of their surprising inability to sell digital advertising in which we were naively dependant upon.
This then forced us to push Adomatic as a mostly consumer facing service. The question for us as a business then became: How can we transform what we’ve learnt from our legacy platform and our more recent advertising investment into a successful product?
As the person championing user-centred design, I needed to also ask: Who are the users and how can this product become something that solves their problems?
The following months would then become a journey into refining the technology into a fully automated service as well as collecting feedback on iterations as early as possible to help steer the product in the right direction.
Adomatic Screen Record — How it works
Wireframe Iterations for the tool page
Facebook Campaign Artwork
Email Campaign Animations
Still to come…
- Rationale in the design decisions for the above sections
- How early prototypes were developed and what was learned from them
- Designing an animation engine using computer vision and OCR metadata
- Transferring human designer problem solving skills into computer algorithms
- Building an editor and not letting it unbalance the concept of automation
- Creating a semantic model with the prospect of creating a domain specific language
- Using UserTesing, Peek and Live Chat to collect qualitative data on a shoestring budget
- Using Mixpanel and Lucky Orange to collect quantitative data
- Building concierge MVPs to test different verticals and justify product development