Hi my name is Pratik Zaveri.
I am a user experience designer with over 6 years of experience building consumer, mobile and enterprise applications. I have a Masters degree in Human Computer Interaction from Georgia Tech and an undergraduate degree in Electronics engineering. I am passionate about solving complex problems and enjoy working in environments that are team-oriented and mission driven. I’ve worked on a wide range of projects, from a award winning education application to a highly complex advertising platform. This portfolio highlights some of these projects.
I am more than happy to deep dive into any specific project upon request and answer any questions you might have.
A few examples I would like to show you
- Yahoo’s Demand Side Platform
- Analytics, Reports & Dashboards
- Enable Mobile - Learning App and Authoring Tool
Yahoo - Demand Side Platform
What’s a Demand Side Platform?
Simply put, a demand side platform (DSP) is a piece of software that digital advertisers use to find and target audiences on the web. A DSP features something called as Real Time Bidding, which works like a stock exchange - evaluating offers or bids in this case and serving the ‘winning’ advert.
Versions and Timeline
Summary slide - What it is and why is it important
The internet is powered by ad revenue. Every large internet company is basically an engine to gather information about their users - everything from movies you like, shows you watch, places you shop at, political leanings, your travel plans and so on. The richer the data the better the change at a ‘conversion’ i.e. getting the user to click on an ad.
Back when I joined Yahoo in 2011, the primary form of leveraging user data to serve ads was through private deals with advertisers. This is called Premium or C1 advertising. Although profitable, this really restricts the scale at which this data can be leveraged. There are only so many sales people one can hire. The solution - create a platform that our end users can access - give them the tools to identify users that they want to target and where they want to target them - make money off of the clicks or conversion. Simple.
The Mission - how did we build this?
My first project at Yahoo was to understand the advertising ecosystem (this diagram will help clear things up!) and to build Yahoo’s first DSP. A pretty non trivial problem to solve right out of grad school.
My Role and Responsibilities
I was the first UX designer on the team that eventually grew to 8 designers - including both interaction and visual designers. //I’ve said designers a lot.
My key contributions over the last 4 years (up until our most recent iteration) have been around understanding and optimizing key workflows and feature sets - from campaign planning, creation, optimization and reporting. My deliverables are usually wireframes that illustrate the user flows and detailed interaction designs. I usually create prototypes in Invision or other similar tools and can deliver pixel perfect mocks when required.
Process (typical) - Since it has evolved over time
Our design process is very team oriented and involves close interaction product mangers, researchers and designers. It is also very user centered as every iteration is tested with a pool of users across a variety of roles.
The advertising ecosystem
To better understand the digital ad industry I read a number of papers published on the subject, attended internal training sessions and set up a ton of 1:1 with SME’s and internal users. I created a glossary during this time to capture the endless acronyms floating around and would update it with every new term I learnt. This glossary actually became a useful resource for people who joined after as a sort of starting guide.
Step 1 -
I worked closely with a product manager and a team of subject matter experts to whiteboard key workflows, feature sets andmajor pain points. The subject matter experts in this case were from a mobile DSP company acquired by Yahoo around this time. I also did a comprehensive competitor evaluation of market leaders in the space like Google’s DoubleClick, Turn, MediaMath and Tradedesk.
Some key workflows that we identified at this stage and that we needed to build to achieve feature parity with the market leaders were -
Audience building and Targeting
Creative assignment and Trafficking
Step 2 -
Generative User Research
After this initial data dump of ideas and issues, the next step was to talk to some internal users to better understand what their expectations were from a tool like this. Also important was to get a better sense of how this product would impact the traditional model of selling premium advertising, if any. Most critical though was the need to get a thorough understanding of the different roles within the company that are involved in the process.
Our resident researcher did some fantastic work to help us better understand the lay of the land.
Example of John or Joanne slide here
Generating ideas for design concepts
Armed with this knowledge, the designers (we had one more designer join the team at this time) started working on design concepts for the product. The approach was two-pronged. First was the actual feature work for the key workflows and the second was to identify and create standards for multiple and reusable components like navigation, grids, forms etc.
For the first part, we worked closely with the product manager who managed to distill all of the information we received from the generative research and competitor evaluation into crisp user stories, occasionally with some acceptance criteria thrown in!
We worked on a two week sprint cycle with check ins twice a week with the product org and weekly check ins with our end users. The weekly sessions with the end users were incredibly beneficial, helping us stay on course and provided us with invaluable design feedback. .
My approach was to create multiple design variations for a specific problem and to have users comment on their preferred approach and provide recommendations. Every weekly check in, I would present at least 2-3 variants and get some great feedback.
After the first 1.5 months or so, we had some standards in place around the key components of the system in place that could be leveraged for a variety of use cases and sped up the development considerably.
We had a large and very talented engineering team comprising of around 12-14 engineers working on this product and we shipped out theAlpha version product in just over 1 quarter.
We generated over 1 million dollars in revenue in our very first quarter after release.
Vinnie quote here
Version 2 and Beyond
The first version of the DSP was released in early 2014 and has been completely redesigned twice already. I’ve been a integral part of the team throughout, using insights I learnt when building the first version to help guide design for future releases.
Yahoo acquired (acqui-hired?) another ad company called Bread around this time and we had another 4-5 designers from Bread join our team. I worked closely with most of them and brought them up to speed on the products, the working cadence and the vision for the next iteration of the product which were difficult to incorporate in the first version. I’ve always enjoyed working closely with teams that are acquired by Yahoo, as designers from a smaller company wear multiple hats and bring a lot to the table. I’ve since worked closely with the designers from BrightRoll (acquired 2015) and most recently with designers from AOL.
Design Standards -
For the second version of the DSP, I was involved in the design and documentation of a formal style guide among other things. The scope of this style guide was not limited to the DSP but was intended to bring all the different ad products under a single design umbrella. We supported over 20 different advertising products at this time and coming up with a group of components that would be reusable and scale for a wide plethora of use cases was a challenging task.
A screenshot of design style guide
IA and workflow improvements
There was a significant upgrade in both the workflows and the visual aesthetic of the product at this time, as we had more designers and a lot more time to work on it.
Specific features and some screen shots from this second evolution of the DSP.
Q- Should I speak about the bulk editor here?
---screens go here---
The DSP Today
I am currently involved the third iteration which is also the most ambitious of the lot. Codenamed DSP Prime, the goal is to rethink all of our primary workflows to made the experience more fluid and interconnected.
Design process updates
We’ve also made significant updates to our ways-of-working and have been trying to incorporate
More in-depth heuristic evaluation
A good example to illustrate how I am thinking about this is to take a look at the updated Line Setup workflow.
def - A line item is a construct within which an advertiser can decide which audience they want to target and how much budget are they willing to commit. They can then set a date range to serve ads to those users.
I’ve designed the revamped Line Item workflow with
Pick one aspect to dive deeper into - (Line Item Setup E2E) - Keep this easily digestible.
Mini case study with research design around it
Other areas of involvement - feature wise and design stage wise - Link to a detailed case study on different facets of the design.
Design MentorOver the course of the last 4 years, to new designers on the team. I make it a point to get well versed with the industry and all surrounding facets before starting to put pen to paper. It helps when having conversations with product and understand why they are recommending a particular feature or workflow. Having context of how the pieces fit together is invaluable for the actual design.
Goal is to release an alpha version to external users by the end of the year 2017.
New Targeting Flow
Totally kick ass