Build better product by validating ideas early— the Freshworks Way

The UX behind delivering exceptional customer support at Freshworks (formerly Freshdesk).

I am Ajith Sowndararajan, UX Manager at Freshworks (formerly Freshdesk). Back in July 2014 when I joined Freshworks, the company was poised for the kind of growth that most startups dream of. The company had just raised an investment of $31M (series D round) from CapitalG (formerly Google Capital) and had a user count of 20,000+. But, the towering challenge was structuring the design process and UX research.

Me and my amazing friends from work!

Since then we have grown tremendously as a company and a design + research team. Keeping my fingers crossed, with a solid design process and the right collaborative tools, we are aiming to grow to new heights.

Today, in this article I will talk about the process of user research and testing we follow at Freshdesk, how this process has evolved over the years and how CanvasFlip is helping us make smarter and faster design decisions.

Evolution of design process in global SaaS teams

Design teams at Freshdesk try not to involve the developers to build a HTML in the early stage of the project, a practice that used to be the norm few years back.

Retrospecting the traditional design process : Few years back, once the concept was validated, designers came up with the solution as images or PDFs, which the developers coded instantly. The usability testing was done after the HTML was built and the iterations followed. This obviously killed a LOT of time and required additional development efforts.

This process is totally disrupted today. It’s more AGILE now.

Today the solution is validated with a clickable prototype. We iterate on the solution until we agree on the basic flow.

I would say that every team should validate an idea before development team has given enormous time to it. It’s a huge save on time and efforts. They need not rework on a lot of things.

Before you start solving, make sure that the need exists.

The first step in building a product is to feel the pulse of your target segment and identify the need in that segment.

If the problem you choose to work on doesn’t actually exist or the solution you propose doesn’t actually solve anything — It does not quite excite the user to adopt the solution or the product.

That’s why I validate if the need exists, even before I start thinking of the solution. The best way to do it is to roll out a poll on your Facebook or Twitter handles or send out a survey to groups which has people similar to your user persona and see if the need exist.

There are a range of tools that you can start off with for the poll/survey. Facebook surveys, Typeform and Google forms being my personal favourites.

The need exists. Now let’s validate the solution ;)

The process of validating the solution at Freshworks has evolved over the the last year. We have always been very passionate about user testing and tried to grasp what the user needs. But then, what has changed in the last few months is — “We chose CanvasFlip for quantitative data on our designs, as well.”

Before CanvasFlip, here’s how we went about validating solutions -

  1. Build a basic HTML prototype
  2. Shared this HTML prototype with our test users, or gave users the control of our browsers,
  3. Once users were on it, we monitored how they use the prototype
  4. Also, to hear them think aloud, we connected over GotoMeeting or Hangout.
  5. With all this exercise we gathered feedback and tried to translate them into design iterations.

Honestly, with this approach, it was very difficult to VALIDATE a concept.

The challenges

1. Catching users in their comfortable time
While building a global SaaS product from India, we also essentially have to validate with users across the globe. They can be in US, Europe, Southeast Asia or absolutely anywhere. Part of my job is to be the user’s voice. But, it often became a challenge to sync time with users. Users naturally interacted with the solution in their free time, sometimes it was 1.00 am IST. Anyone who is into the same business would resonate with my thoughts about how difficult it is to match timings with users.

This process of validation was kind of slowing us down. A huge chunk of my working hours went into scheduling these tests and adjusting my own calendar every time when a user reschedules a call..

2. Difficulty in drawing quantitative insights from just watching users
We were using multiple tools for parts and pieces of our workflow. Apart from the hustle in using so many of them, we weren’t able to get quantitative data such as a heat map or user flow. With this process, it was difficult to observe an interaction pattern of problems that users were facing.

Without clear insights it was difficult to proceed with a design and objectively validate it.


All these challenges impeded our ability to efficiently validate if the solution worked or not.

Thus began the hunt for a new tool that could solve our challenges. After hearing good things about CanvasFlip, we decided to take a closer look.

Taking a closer look at CanvasFlip

First time I heard about CanvasFlip from one of my colleagues, I checked out their website. It said, “User testing on prototypes”. That instantly caught my attention. But once I started using the product, I liked what all they had to offer. There was so much I could do, starting from prototyping to setting up an entire user test, gathering user videos and other quantitative results.

This is what I had wanted all this while!

Validating a “solution” with the tool

Adding CanvasFlip to our workflow, not just made user testing easier but also more efficient. Here’s how our workflow remoulded with CanvasFlip.

1. Prototype creation and sharing with a closed group

Working aligned to the user’s need, we create a prototype of the proposed solution. To start off with, we just grab the social share link (the magic URL) and share it directly to a closed set of audience, over emails. But other social media channels like, Twitter, Facebook, Slack could be a good platform to share as well.

Creating a prototype has always been the starting point for me, it’s just that magic URL is making my life simpler now :)

2. Setting the context of user testing

It’s important to show a trailer before you launch a superb movie! Similarly, before users go into user testing, it is important to narrate a story and set a context to them for trying out the prototype. That’s when we get the most fruitful feedback.

Let’s take an example of this.

Say, I’m building an app for the auto drivers in Bangalore that would help them accept rides and go pick passengers from their locations.

I have a prototype in hand which I want to test with the auto drivers. I really doubt how fruitful the response would be if I don’t narrate the scenario (or prologue) to him. Here’s what I would tell him -

You are driving a share-auto in the Kormangala area and you’re heading towards Jaya nagar. You want to pick up passengers on the way. Here’s an app that helps you do that. Will you use this app for it ? If YES, then how?

3. Find out what’s happening quantitatively

Quantitative research such as heat maps and conversion funnels do an excellent job at answering “how many” and “how much” type of questions. An initial analysis of numbers is required to set the conversation with the users, while trying to understand their mindset. This part was taken care of with the CanvasFlip tool. Heatmaps, conversion ratio, time taken for a task, interactions performed etc are easy to collect.

4. Talk to your users and understand “Why is it happening”

Following the quantitative analysis, we make sure we talk to at least 5 of the users to understand things like “where did they get stuck”, “what was their initial problem” and “does this prototype even solve their problem?”… Qualitative methods are much better suited for answering questions about “why”?

This is a very important step because referring to only quantitative data could be misleading. Maybe the user didn’t even get the context right while trying out the prototype..

5. Cross-check and confirm user feedback with analytics

Based on the learnings from user’s feedback, I go back to user videos & flow maps recorded on CanvasFlip, watch their sessions - screen after screen, and try to figure out where exactly they faced friction. Fellow researchers would agree with me when I say, “Most of the times, users say they want something but they keep fiddling with the interface to find a particular information or even a button.”

This is when a combination of quantitative and qualitative research helps.


Letting user feedback shape the experience

At the end of the these tests, I segregate my learnings into two sections -

  • Tactical fixes — Which are at surface level and quick changes that enhances the UX (mostly UI elements) quite a bit.
  • Fundamental or long-term changes — The flow is mapped to the mental model of the users or not

With these user behaviour data, I’m poised about the iterations on the basic design structure/skeleton and I know we are on the right track!

How we solved some of our challenges with CanvasFlip

  1. My biggest pain point in user testing was solved — Users can now review the prototypes on their convenience and I get to analyse it on my convenience.
  2. With a combination of qualitative and quantitative data, I can make a well calculated guess whether or not the concept will work. I need just an evening’s time to comment on such questions!
  3. Validating the idea even before it was implemented. It saves humungous amount of time, efforts and above all helps us stay aligned to what the users need and not what we think users want.
  4. It is quite useful to show value of the concept. It’s easier to convince other stakeholders and team members about the concept when you have both quantitative and qualitative data.
  5. Having developers on the same page saves us from landing into communication gaps. Same with colours, opacity and everything else. CanvasFlip also has the capability to handle SPECs ( the design hand-off module) to handle entire design workflow.

Conclusion

One of the main purpose of design is communication. My job is to set this communication between the product and the user — Right. In the process, a designer has to be very careful of what he/she is creating and communicating. At the same time, it’s important that the team believes and is convinced about the power of the solution. I believe, design is both subjective and objective — It’s about how you present it.

“Tools like this gives designers a voice, backed up by both subjective and objective data.”

Fellow Designers and UX researchers, go create an amazing solution that your users need and approve of! (And ya, these guys at CanvasFlip are always enthusiastic in helping you do that! )