Thoughts from my journey

just be …

5 Traits of User Onboarding Craftsman

By Guy Nirpaz August 28, 2012

If you haven’t looked at Jordan Koschei‘s analysis of dropbox user onboarding flow, you should go ahead and read it here. I’ve also found this interesting thread on Quora with other examples of good user onboarding flows.

User onboarding is a critical part of the product usage and adoption lifecycle. When users have the option from thousands of products and services to choose from, the first ‘feel’ of a product becomes very important to the success of a new product or service implementation.

Every online product manager is aware that to achieve both, there is tension between:

getting people to complete necessary steps in order to start seeing value from their product; and at the same time
making sure not to overwhelm them with unnecessary information/tasks/steps.

In the case of Totango onboarding, we’ve had a heated internal debate with one camp feeling that we’re imposing too much effort upfront by requiring a Javascript implementation, and the other camp feeling that we could ask for way less information and provide just enough value before getting deeper. We’ve called the new approach the ‘drop-in’ project. I’m happy to admit, as the leader of the ‘aggressive camp’ that I was completely wrong. The drop-in yielded fantastic results of higher rates and much faster onboarding.

Crafting a great user experience while building a successful user onboarding flows is like an art, or more precisely a craft of minimalism.

Too little effort upfront will leave users underwhelmed with your product and too much effort upfront will turn away new users.

I’ve been thinking a lot about this problem and have been working with Totango customers on their user onboarding flows. You can read about nCircle’s journey here. From working with many customers like this, I have been able to find 5 rules of thumb that lead to great user onboarding implementations:

1. It’s all about the core value
When designing the onboarding implementation you should carefully understand the core of your application. In most applications, there are many things people can do, however, you’re designing the best way for people to understand your service value without interruptions and without extra unnecessary steps.

2. User-centric approach
Think about who your user is in a very concrete way. Where did they come from, what have they done prior to initiating your “getting started” experience, what language do they speak, which device are they using? By preemptively getting this data, you don’t need to ask redundant questions and they don’t need to fill data you already have.

3. Have an owner, have a team
Someone has to own the onboarding project, which will take time to get right. Have someone (you?) own it and drive it to success. It will require many follow up changes and decisions that will impact the results. Also, important, make sure you have an implementation team in place that can follow up with UI, graphics, flow, copy (language) changes.

4. Iterate, iterate, iterate…
You will probably need hundreds of changes, some of them very minor while other very significant to get it right. You’ll have to go through many cycles of analysis and learning to get it right.
The key metrics you should care about are:

Onboarding success rate – of the people who started (and you care about) how many successfully completed the onboarding flow, and how many dropped and why?
Time to onboard – optimize for the minimal time to successfully onboard new users. Track that time and find ways to improve it.
Engagement – the rate of onboarded users who ended up frequently using the service/app
Monetization – the rate of successfully onboarded users who ended up paying

5. Remove steps

“A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away“. ~Antoine de Saint-Exupery
This is an all-time favorite of mine, and it’s true in user onboarding as well. Always consider removing steps, you’ll be amazed by the results.

Which products do you find with the best onboarding experience? Let’s use the comments thread below to praise those products.

Best Practices , churn , Customer Engagement , customer lifecycle , metric , User Engagement , user lifecycle.

3 Steps to Measuring User Engagement with your Web Application

By Guy Nirpaz December 26, 2011

Take a step back and you’ll realize user-engagement is the single most important metric in any SaaS business.

In a world where customer evaluate your offering at their own pace, and can cancel their subscription at any time, the best way to maximize your business potential is to make sure users are engaged and see value in your offering.

And the best way to ensure that is to create a metric which can be monitored for change and improvement on an on-going basis.

It has been somewhat surprising for us to see that most SaaS companies (and we’ve spoken to hundreds by now) are somewhat at a loss as to how to actually measure their user’s engagement with their offering and application. In fact, when pressed many admit that, important as it is to their business, they actually *don’t* measure user engagement. Simply because they could not figure out a systematic way to!

Since we’re here to help SaaS companies do better in this area, here’s our 3 step guide to getting started.

Step 1: Segment your users into lifecycle stages

The signals engagement for a trial user that has just signed up vs. an established customer is very different. Trying to come up with an engagment metric that applies throughout all lifecycles is practically impossible. Consider the following (in a fictitious SaaS application):

User1: Signed up last week, has logged in 5 times, created a project with some content and reviewed our knowledge-base 3 times
User2: Paying customer for a year. Last week logged in 5 times, created a project with some content and reviewed our knowledge-base 3 times

Clearly User1, as a new trial user, is exhibiting a good level of engagement, where-as the behavior of User2, a year into their subscription, is concerning at best.

We recommend you break down user lifecycle at least into the following stages. We also suggest some ideas of things you’d want to look at as you compute engagement at each stage

You’d want to apply a different engagement metric to users depending on where they are in the process.

Step 2: Create a scale

Engagement is not a binary metric. Users are not either engaged or unengaged, but rather fluctuate on a scale. We recommend creating the following buckets:

The time window to measure varies. we typically suggest 14 days – 30 days, depending on the application’s complexity.

For a top-line view, you eventually want to end up with a dashboard similar to that shown below.

The chart shows, number of total, highly engaged and lightly engaged users overtime. For convenience, we overlay important milestones (product releases, marketing campaigns, etc.) so we can see their affect on our users.

For example, we see a good pickup of total activated users after launch . Growth is mainly in lightly engaged users however.
Important milestone 1 made almost no impact (maybe it wasn’t that important after all?
Important milestone 2 on the other had, clearly increased the number of highly engaged users (we should do more of that)

Step 3: Constantly refine & improve

Your engagement metrics should not be static but evolve over time. You should constantly “test” them against users eventual decision to purchase or cancel their subscription. If they don’t provide a good enough prediction as to what a user is likely to do with their account, the metric and its underlying formula should be tweeked.

Specific Examples:

“Highly Engaged” Trial users should convert at a very high conversion-rate to paying accounts
“Gone” and “Fading” Paying users would tend to churn if left unchecked