Troubles and Triumphs When Re-Building Our Customer Onboarding Process at WordPress.com

I’m on stage right now at SupConf ATL talking about some work we’ve been doing recently over at WordPress.com to rebuild the onboarding process for our Business plan customers. What you’re reading below is a transcript of the talk. I’ll update this page with a video whenever it’s available! Full slides are at the bottom.


Let’s run through a hypothetical scenario for a second. Imagine you’re looking at the refund and churn rates for customers at various levels within your product. Now, let’s say that you notice your highest paying customers were also at the top of another category – highest refund and churn rates.

What would you do?

That’s the exact scenario we found ourselves in over at WordPress.com during the middle of 2016. Our highest paying customers on the Business plan were refunding at an alarmingly high rate compared to customers at other plan levels. We had both a challenge and opportunity ahead. If we could somehow solve a piece of this problem using Support, we could directly contribute to the bottom line of the business in a meaningful way.

Fast forward to the present day, we haven’t solved the problem entirely just yet. In fact, I don’t think it’s a problem we’ll ever be able to 100% solve. Despite our best intentions, someone will inevitably refund for one reason or another. However, comparing our refund rates for these customers from 2016 to 2017, we were able to reduce the refund rate by over 30% while also seeing an increase in total signups. This represents a significant impact in the bottom line of the business!

A lot of factors are at play here; I won’t pretend that Support is solely responsible for the boost. Our team launched a really successful update to the Business plan allowing outside plugin and theme uploads. Our interface continues to improve. There are a lot of moving parts. However, during that same time, we completely revamped our approach to onboarding customers within WordPress.com support.

Today, I’d like to share our approach with you including how we identified problems we could specifically tackle within support, how we launched new onboarding initiatives for all of our Business customers, and ultimately where we’re at today with our customer onboarding. Along the way, we’ll talk about successes, learning points, and questions we’re still trying to answer.

Understanding the issue you’re trying to solve

Customer onboarding has become a bit of a buzzword over the past year. There’s a ton of fantastic information and specific examples available now from companies like Highrise, ConvertKit, Intercom, Helpscout, and many more. Each company offers a unique approach. Highrise and ConvertKit, for example, send personalized videos for every new signup. If you sign up for Wistia, you’ll get this awesome video of Jeff Vincent showing you how to use the product.

All of these approaches are fantastic, but I’d like to take us back a step – back from the actual implementation. Before we start talking about tools and tactics, we have to understand the problem we’re actually trying to solve. This has two key components:

  1. We can clearly define the problem; we know exactly what it looks like in the customer journey.
  2. We are in a unique position to impact the problem.

If we can’t articulate a problem that meets both of those criteria, we’re just going to be chasing our tail if we start to think about tools, videos, email series, etc.

Let’s take the problem we were facing over at Automattic: Our highest paying customers were also churning at a high rate.

That left us with several questions including:

Why were these customers churning so quickly and not those at a lower price level?

What were we missing that these customers were looking for?

To find out the answers, we took several different approaches. First, we put a survey in place for every single refund and cancellation. Second, for a period of 30 days, we emailed every single signup for the Business plan and offered to sit down with them for 30-45 minutes. During that session, we helped them with their site, but we also asked questions like “What were you looking for when you came to WordPress.com?” and “What kind of expectations did you have when you signed up?”

We collected months of data from cancellation and refund surveys. When we combined that data with our individual interviews, the problem we were trying to solve began to get a bit more specific.

One group of customers was refunding because they couldn’t install plugins and themes at WordPress.com like they could on a self-hosted site.

Another group of customers was refunding because they were looking for a drag and drop interface.

A final major group of customers were refunding because it was too hard to get their theme/site looking just how they wanted.

From a support perspective, we couldn’t directly impact the first two elements (although both elements are currently or have already been addressed). However, we were in a unique position to address the third element. That was our job – help our users to experience success using our product.

The point I want to drive home here is not that you should contact all of your customers for a month and talk to them individually. That may or may not be the right approach for you. My point is that you should look to understand the exact issue you’re trying to solve on a deeper level. In our case, the problem shifted from:

Our Business customers are refunding at a high rate.

To:

Our Business customers are finding it difficult to build a site and set up a theme to match their vision.

That hits both of our criteria mentioned above. I know exactly what it looks like – activating multiple themes, reading through theme documentation, and enviously admiring sites I enjoy – and we’re in a unique position to have an impact. The next step is defining exactly how we’re going to make that impact.

Defining the onboarding methods that match your product

As mentioned above, there are a lot of ways to make an impact on the onboarding experience for new customers from a support perspective. You can send targeted messages through Intercom or custom videos from an app like Bonjoro. You could send custom gift baskets or default to emails containing more general information.

There isn’t a one-size-fits-all approach, but I wanted to propose a framework for categorizing these onboarding mediums that I find helpful. Then, we can walk through how this ended up applying to Automattic.

First, let’s pretend we have two factors to with:

Scale – As your business grows, how easy is it to scale up this method of contact?
Personalization – How individualistic is this approach?

These factors are inversely correlated meaning as one goes up, the other typically goes down. This leaves us with three “segments” to operate in.

High Personalization/Low Scale
This would be something like individual videos or handwritten notes sent to every new customer. I categorize our individual calls with new customers here as well. The benefits here are pretty clear – when delivered correctly, you can deliver a personalized experience that feels authentic and is very hard to replicate.

Moderate Personalization/Moderate Scale
Webinars fall into this category. In-person device classes at Apple would fall into this category as well. When delivered correctly, they can still hit on the personalization piece, but at the same time, your customer knows that they aren’t the only one in the room. On the flip side, this scales relatively well.

Low Personalization/High Scale
The traditional guided tour or onboarding email campaign falls into this category. It’s the kind that we’re all most familiar with and the easiest to implement. Virtually every service you sign up for sends some kind of welcome email with a tutorial or follow-up series walking you through setup. There are definitely ways to still incorporate a bit of personalization here. I think Nathan Kontny and the Highrise team have a really interesting approach, for example, but compared with the other methods, this offers the lowest levels of personalization.

Now that we have our three segments laid out, which one should we shoot for?

It’s very tempting to go for the high personalization/low scale option. Frequently, that offers the least amount of competition and results in the highest level of amazement from your customers. But, I actually want to posit a different approach. Each of the segments above has its own pros and cons. Harkening back to our well-defined problem from above, there isn’t a catch-all answer to which segment best fits every problem.

Let’s take a closer look at the problem we’re trying to solve and pull out which segment(s) might best apply:

Our Business customers are finding it difficult to build a site and set up a theme to match their vision.

First, we already have support documentation in place. We could spend extra effort doubling down on the high scale solution, but we’re likely to find ourselves back in the same place with little to show for our efforts.

Second, it’s unlikely that many Business customers share the same vision for what they want their site looking like. They each have unique needs, preferences, likes, etc.

Third, WordPress can historically be difficult to set up. I’m not sure if you have used it to setup a website (if not, let’s talk!), but when you’re brand new, it’s tough to understand widgets, themes, custom CSS, etc.

There isn’t an easy rule that applies in all cases, but in my mind, low personalization/high scale options work well if you have a product with few variations. It’s great at doing one thing, and the path to that one thing is very clear. On the flip side, high personalization/low scale is appropriate if your product offers tons of customization. Each customer might come in with a different use case and perspective.

Excelling at onboarding that doesn’t scale well

In our case, each customer journey is so unique that a high personalization experience is a great fit. Now, how do you solve for the scale piece?

Let me first describe the current process we have in place for handling these concierge sessions.

We invite every single new Business customer to a one-on-one concierge session with a Happiness Engineer immediately upon signup. We send out this offer through a few different methods – immediately via email with the receipt, in-app on post-purchase screens, and via WordPress.com notifications. The customer can then click on the link provided, see our availability (managed through Calendly), and schedule a session at a time that works for them.

The sessions usually last between 30 and 45 minutes, and they’re generally handled over screenshare using Zoom. Typically, the customer will “drive” meaning they control the screen, and the Happiness Engineer directs them with verbal cues. During a session, the Happiness Engineer might help them set up an online store, build their theme, or optimize their site for search engines.

From a customer perspective, the sessions are amazing. You’re able to speak with a real human somewhere in the world (Automattic is all remote) and ask any question that comes to mind. Compared with other methods of support like live chat and email, we’re able to get a lot more done in those 30-minute sessions with the voice component.

From a Happiness Engineer perspective, the sessions are equally valuable. In support, we often connect with our users in a text-only environment. This screenshare environment provides a unique opportunity to actually talk with our users one-on-one and understand their issues and needs. The sessions can also be challenging as they require pretty extensive knowledge across our product lines.

Obviously, these sessions are time-intensive and require a large investment of our people resources. We have made a few trade offs that will hopefully allow us to continue scaling the number of sessions we’re able to complete in a given month.

First, we only offer the sessions in English. We provide support in Brazilian Portuguese and Spanish over live chat and email, but the logistics of matching a user up with someone that speaks that language for a screenshare proved too difficult for right now. It’s certainly something we’re hoping to offer in the future though.

Second, we use Calendly as the canonical place for scheduling sessions. This means that if a user needs to reschedule, we cancel the session, and they have to go back and find another time. Similarly, if a user wants to set up a follow-up session, they have to use Calendly as opposed to working directly with a Happiness Engineer to find a time that works.

While these two choices are inconvenient and sub-optimal from a customer perspective, they have allowed us to move a bit quicker and reach more customers overall.

Let’s circle back to our problem mentioned above and the obvious question:

How effective have these sessions been in helping Business customers build sessions from scratch?

We’ve been able to complete thousands of these sessions over the past year, and the results thus far indicate that customers that complete a concierge session are half as likely to refund compared with those that don’t. As I mentioned above, there are other factors and feature releases at play here, but we’re still really excited about the results we have seen thus far.

Alongside the payoffs in refund and churn rates, these concierge sessions have led to many other benefits. For example, after each session, the Happiness Engineer jots down some quick notes including which topics were discussed, what goals did the customer have going into the session, etc. This helps our NUX team to develop a better understanding of our business customers.

Second, many of our developers and designers are performing ride-alongs meaning they’ll listen in to a session to get real user feedback and again develop a better understanding of our user base.

We still have a long way to go with these sessions. The scheduling aspect is still very manual; we’re still battling a high no show rate; and, we need to build more processes, procedures, and resources to allow new Happiness Engineers to get involved. The progress we have made in the past year has been inspiring to see though.

Onboarding can take a variety of forms. What has worked for WordPress.com and Automattic might not work for you. Your mileage surely may vary. However, by first understanding the problem then picking the right medium for your audience, you’ll experience more success than if you were to throw things at the wall and see what sticks. This thinking process ultimately led to an initiative that has had a significant impact on our bottom line, and we’re incredibly excited where it can head in the future!

Full Slides

One thought on “Troubles and Triumphs When Re-Building Our Customer Onboarding Process at WordPress.com

Leave a Thought