Making people fill forms

Some insight from our experience on one of the most frustrating areas in mobile UX.

One of our products, is a sales enablement platform. And one of it’s core value propositions is that salespeople actually use these apps, even when on the move or in front of a customer.

This is far easier said than done. Making a field sales rep use a new app is like making my infant daughter sleep at night :) — it requires intimate knowledge of the target’s pressures and levers, innovative solutions, and multiple skills to execute these innovations.

I’m detailing one of these innovations today, in the hope that it will help some of you.


Our customers typically have a large suite of complex, often intangible products, sold by field sales personnel who don’t understand most of these products. Think mutual funds sold by bank clerks, or life insurance sold by agents. So matching their prospects’ needs to a particular product is a big challenge.

The traditional solution is a “Product Selector” or “Need Analyser” sheet, where you enter a few details, and get a list of matching products. We started off with one such form: 5 simple fields to enter, and immediate clarity on what to pitch.

Needless to say, it didn’t work. Users don’t want to fill in even 5 fields. In front of the customer, even 60 seconds spent looking down at a smartphone, were a deal-breaker.

In front of the customer, even 60 seconds spent looking down at a smartphone, were a deal-breaker.


So we dug deeper and, among other things, spent a night poring over videos of our users meeting clients. The issues got clearer:

  1. It cost too much time to enter text or select from drop-downs
  2. Users couldn’t quickly change a parameter or two, and experiment with the result
  3. Validation was a pain, both for initial entry, and for change experiments


We then tried multiple input formats, with these 3 objectives:

  1. Make it easier to select/enter parameters
  2. Make it impossible to skip parameters, or give invalid input
  3. Make it easier to play around with options

We tried out various solutions, ranging from chat-based entry to voice recognition/dictation. Finally, a simple solution tested well:

  • Step-by-step entry (we serial-ized a usually parallel form)
  • Customized input formats that ensured speed and eliminated errors
  • A structure that encouraged quick experimentation
  • An 8-second cycle to enter 5 well-considered parameters

Sounds too good to be true? Do take a look at the solution below.

An example with mutual funds in an Indian context — note the “Child’s Wedding” goal :)

This tool, in it’s very first deployment, is being used more than 20,000 times a month, by agents of our first client alone.

The lessons we learnt

Great inputs can create large deltas in usage. So when the feature is central to our user’s needs, it is worth spending these 3–4 days on input. Every input field is unique, and can be made easy, intuitive, and auto-valid with a little effort.

On a side note, when serial-izing a parallel process:

  1. Input has to be easy, ideally pre-validated — think drop downs vs text input
  2. At any point, users should be able to trace back their selections/inputs — think breadcrumbs or step-counters
  3. If you have more than 5–6 steps, it’s too much for one continuous serial input — find a logical split

I hope this helps you generate ideas, or validate what you are doing. Have your own experience to share? Disagree with something above? Do drop me a line.