IA & Content for an Ed-Tech Startup

LIGHTSIDE

THE GOAL

Re-architect the website for a rising startup. Replace the old structure — the design of which had been cobbled together on assumptions and content that were increasingly outdated — with a more intuitive one that can be built upon by future designers as the business evolves.

Additionally, plan for and develop content that explains the startup's complex product — machine learning software for education.

What I Did

Stakeholder Interviews, Content Inventory and Audit, Proto-Personas, Competitive Analysis, Affinity Diagramming, User Flow, Site Mapping, Requirement Elicitation, Task Analysis, Sketching, Wireframing, Content templates, Copywriting, Branding, Channel Strategy, Documentation

Background

This work was done during the summer of 2013, when I was an intern for the education-technology startup LightSide.


THE QUEST

Everything is better with affinity diagramming.

  • Interviewed stakeholders within the company to assess needs, set goals, and elicit requirements.
  • Mapped the existing architecture of the website, alongside a detailed inventory of its content, and followed this up with an evaluative audit to figure out what was still useful, what wasn't, and where there were gaps that needed to be filled.
 

The final architecture. Lean, clear, and extensible.

BUMP IN THE ROAD!

Who are the users? As a startup, LightSide was actually still figuring that out. I made do with a collective workaround of low-fi personas, competitive analysis, user flows, and additional interviews with stakeholders.

A clear and simple architecture began to emerge through synthesis of all the research, brainstorming, and affinity diagramming.

 

 

  • Narrowed down content needs (drawing from business goals and personas) and decided which formats would be most effective and appropriate for each.
  • Created content templates to guide and coordinate development, making explicit the purpose and audience for each element.

Again, though, what did I really know about my audience? Because LightSide was an early stage startup, it was still in flux who exactly this content was for. The trick was to leverage what I did know, and to go where the stakeholders were — comment boards, articles, op-eds, and online forums were full of people in dialogue about similar products and competitors. This was a trove of insight into the minds of supporters and critics alike.

 

 
  • Identified current and potential channels by which LightSide’s various audiences would engage with its content.

  • Prioritized channels based on a cost-benefit analysis of each and an inward-looking consideration of what the company had the resources to do.

  • Illustrated how the channels were related and could be managed going forward, making recommendations for future development and documenting everything that had been done in the content creation process.


THE RESULTS

LightSide's new website architecture was instantly more intuitive, clean and adaptable, built to fit its audience and satisfy the needs of both users and stakeholders. Now a year later, the look and content of the site has been revamped ... but for the most part, the same architecture undergirds its ongoing growth and development.

This work also resulted in content that explained the company clearly and told its story compellingly, with copywriting carefully chosen to convey LightSide's developing brand across multiple channels.

Here’s an excerpt:

There’s a good chance you’re already familiar with machine learning, the technology that LightSide uses to assess student writing. When your email filters out spam, when Netflix recommends a movie, when Ken Jennings lost in Jeopardy! to IBM’s Watson — all of that involves machine learning.

Machine learning means teaching a computer to recognize patterns so that when you give it new information it’s never seen before, it can draw on previous experience and knowledge. In some sense, this is actually very similar to what people do when faced with new information — we process it and make decisions by relying on memory and experience. LightSide’s algorithms can’t ‘read’ the way people do, but they’re certainly fast learners.
— FROM THE INTRO TO AN FAQ