About
Project indexWe have a spreadsheet that details abandonment reasons for every customer, and I’d estimate we’ve talked to at least half of the customers who quit. The two most common reasons we kept hearing about why sign-up process were: 1. They filled the form out> Clicked submit> It took a while> And an error appeared “Invalid token” 2. Why signing-in using SSO went to the sign-up page. Is there any problem with signing via SSO? But there was almost no data track as a failure in logs, nor the data. Weird. 3. In addition, we had outliers such as people using the wrong email domain, e.g. yahoo.cokm. Given that the quantitative data suggested highly invited members in workspace was a big driver of customer retention, the qualitative and quantitative data told a pretty strong story about what to do to cut down on cancellations. This meant there was a problem with sending user authentication email and those users never have seen registration success page.
https://www.dropbox.com/scl/fi/ov4fx5nvk3epyensv6c06/Pouya-s-Work-samples.pdf?rlkey=xaa2tfhm32q558e5vsc95lfmt&st=0pdb3g1u&dl=0Read more
Read more
How did we start?
https://dribbble.com/poulo3Dribbble
Dribbble
Dribbble
Dribbble
Dribbble
Dribbble
AboutTo better understand the issue, I decided to dive deeper by using Inspect Element and Postman. My goal was to examine the performance of the sign-up button and ensure that it was functioning correctly. I specifically targeted the API endpoint: Endpoint: /api/v1/oauth/signup/email By inspecting the network requests and responses, I was able to track the exact behavior when users clicked the sign-up button. I looked for any potential delays, errors, or failures in the request that could be contributing to the high abandonment rate. This analysis helped me identify whether the problem was with the front-end user experience or if it stemmed from a back-end issue, such as the API not responding as expected. Now, we check the endpoint issue with an engineer and learned it was coming from Segment API which took so much time to response to. And we depended on the process for that.
Here’s what we decided to do based on each of our findings: 1. Moving the Segment API to the background. This ensures it doesn't interfere with the registration process. 2. Improving ACL to limit user invitations to the workspace to users in the user tables only. This way, we ensure that invited users aren't already signed up for the specific workspace. 3. Adding a confirmation field for the email address. This helps users enter the correct email address. 4. Sign-in SSO: We discovered that the reported SSO login issue wasn’t a bug. It was redirecting new visitors to sign up when they tried logging in with a different email. However, we definitely need to improve the UX and messaging here later on. 5. Enhancing mobile responsiveness.
I created pivot tables to compare the differences between current customers and those who abandoned the process. Here’s what I learned about customers who abandoned the process: 1. They were primarily using mobile devices or Safari (based on customer support tickets). 2. A significant number of invited users had already signed up previously in another workspace. 3. A considerable portion of new customers experienced issues with valid tokens, so they never received the correct user authentication email.
About
Abouthttps://www.linkedin.com/in/pouya-l-70531236/The next step was to write an extensive SQL query to extract all the data that could possibly help us answer this question. Here’s some of the relevant data we pulled to expand our analysis and get a broader picture: - The month a customer attempted to sign up - The number of workspaces a customer created each month - The length of the free trial - The number of users invited - The number of potential customers and invited users who never completed the registration In total, we pulled data from 51 different fields.
Project indexCofounder, Product and Design
Analyzing the Raw Data
https://buymeacoffee.com/poulo2https://buymeacoffee.com/poulo2Pulled some data
QPage screenWe started by asking questions. In this case, we were trying to answer this question: “What are the major differences between customers that gave up within the sign-up process, and customers that don’t?” And what are the root cause behind this loss?
Squared the quantitative with the qualitative
Problem
We have a spreadsheet that details abandonment reasons for every customer, and I’d estimate we’ve talked to at least half of the customers who quit. The two most common reasons we kept hearing about why sign-up process were: 1. They filled the form out> Clicked submit> It took a while> And an error appeared “Invalid token” 2. Why signing-in using SSO went to the sign-up page. Is there any problem with signing via SSO? But there was almost no data track as a failure in logs, nor the data. Weird. 3. In addition, we had outliers such as people using the wrong email domain, e.g. yahoo.cokm. Given that the quantitative data suggested highly invited members in workspace was a big driver of customer retention, the qualitative and quantitative data told a pretty strong story about what to do to cut down on cancellations. This meant there was a problem with sending user authentication email and those users never have seen registration success page.
Jan 2021 ~ Mar 2023
Tech Stack: OCR NLu, NLp LLM ReactTypescript Python Laravel Nodejs Gitlab AWS
Squared the quantitative with the qualitative
I was stressed. We were experiencing nearly a 30% abandonment rate in the sign-up process, along with a lower conversion rate than the industry standard, and it was eating me up inside. Each of these potential customers had: 1. Realized they had a problem that QPage could solve (AI workflow automation in their recruitment process). 2. Decided they needed to address that problem. 3. Researched solutions that might solve it. 4. Found QPage compelling enough to try. 5. Advocated for QPage to their team members and stakeholders. After investing so much mental and emotional energy, they decided to give it a try. And then, just within the registration process, they gave up. Argh. We were trying to understand the key differences between the customers who signed up and those who didn't. Initially, we tracked the conversion rate trend month-over-month, but we later decided to analyze it on a week-over-week basis. How did I uncover the issue? Well, it started with a periodic analysis where I compared week-over-week sign-up sessions against the number of actual sign-ups. To expand the research, I checked customer support tickets and landing page chat logs for any indications of this issue, but I found only a few instances. I conducted a small survey within our community, randomly selecting customers from three different cohorts to see if they had experienced any problems during the sign-up or user invitation process. Simultaneously, I used Conversion Crimes to assess the new user experience in the process. (I tested it myself many times and didn’t find any issues). But we still wanted to see what insights we could gain from our data.
Since implementing changes 1-3 two months after, we’ve seen our abandon rate in sign-up process rate drop from 30% to 2% – a 15 folds decrease. We’re going to run another cohort analysis in a couple of months to isolate the impact of each change as it’s still too early to know the long-term impact of these changes, but in the short-term, we were stoked about stopping the hemorrhaging.
Works
Project indexQPage was an B2B SaaS company revolutionizing recruitment with AI-powered solutions. Our platform streamlines hiring with advanced tools like AI Sourcer, Interview Assistant, and Nurturing Campaign Automation, helping recruiters efficiently source, engage, and hire top talent.
The Outcome
Analyzing the Raw Data
https://read.cv/pouloResume
Resume
Resume
Resume
Resume
Resume
I created pivot tables to compare the differences between current customers and those who abandoned the process. Here’s what I learned about customers who abandoned the process: 1. They were primarily using mobile devices or Safari (based on customer support tickets). 2. A significant number of invited users had already signed up previously in another workspace. 3. A considerable portion of new customers experienced issues with valid tokens, so they never received the correct user authentication email.
Acting on the results
DeepOpinion