💡 Problem

1. Limited customization.

   Outdated, generic templates made upsells feel disconnected from a creator’s brand, limiting their ability to create a compelling post-purchase experience and optimize for conversion and AOV.

2. Creating upsells was tedious and repetitive.

   Upsells weren’t reusable. Creators had to manually create a new upsell from an offer each time, making it time-consuming to maintain and scale their upsell strategy.

3. No performance visibility.

   There were no upsell-specific analytics, leaving creators without visibility into conversion or revenue impact and no clear way to understand what was working.

🎨 Key decisions

1. Redesign the upsell template to match the creator’s brand.

   The existing upsell page relied on a rigid, generic template, with limited customization, that made it look disconnected from the rest of a creator’s site. I redesigned the template with updated controls that gave creators meaningful control over the page, including layout, typography, colors, imagery, content, and sections. Rather than adding a long list of one-off settings, we brought the upsell experience into the same flexible Builder model creators already use elsewhere in Kajabi.

🎨 Key decisions

2. Make upsells reusable.

   Previously, creators had to recreate an upsell for each offer. We separated the upsell from the offer so creators could create it once and reuse it across multiple offers. This also meant changes could be made centrally and automatically reflected wherever the upsell was used, making the system easier to maintain and scale.

🎨 Key decisions

3. Give creators visibility into performance.

   An upsell is only useful if creators can understand whether it’s working. We added purchases, conversion rate, and revenue at both the overall and individual upsell level, giving creators a way to measure performance and iterate on their offers.

🎨 Key decisions

4. Call the experience “Upsells”.

   The biggest challenge in this project was landing on the correct verbiage for the upsell creation experience as it relates to Offers. We explored several naming directions, including “Upsells & Downsells” and “Post-Purchase Offers”. After some debate and back-and-forth, we landed on “Upsells” because it was the most familiar term for creators. This simplified the mental model, but meant treating downsells as a use case of the same page rather than a separate page type. The tradeoff that we made is de-emphasizing the distinction between upsells and downsells upfront. We assumed creators would use the same reusable page for both, giving us a more flexible system that could eventually support an upsell funnel builder—where the same page can function as either an upsell or downsell depending on where it appears in the funnel.

🚀 Outcome

• 10.88% conversion rate. More than 1 in 10 presented upsells converted, with a 12.71% conversion rate at the first upsell step and 4.29% at the first downsell step.
• $256K GMV in the first month. The new experience processed $256,237 in GMV within its first month, with an average accepted upsell value of $123 and downsell value of $58.
• 10,639 accounts enabled. The experience reached 10,639 accounts across 15,708 sites, representing 60% of accounts with at least one upsell.
• 28K+ upsells migrated or created. Creators migrated or created 28,233 upsells, with 17,487 actively published.
• Attach rate continued to grow post-launch. Upsell attach rate increased from 2.09% to 2.50% between March and June, a 0.36 percentage-point increase YoY. Combined upsell + downsell attach rate reached 2.63%, up 0.39pp YoY.
• Revenue continued to grow YoY. Upsell GMV reached $8.7M YTD, up 11% YoY, while downsell GMV reached $331K, up 28% YoY.

💭 Reflection

📚 Research

To understand the domain, I initiated a mind-mapping exercise with my PM in December, before we kicked off the project in January. We basically put everything we knew about affiliates onto a whiteboard, which helped me map out the current workflow and identify areas where we needed to do more research.

The journey of setting up the affiliate program looked like this: they start off with affiliate registration, create a landing page to promote their program then qualify them using our survey tool. After the affiliates join the program, they then set up share links and coupons for the affiliates to use. They would announce via Kajabi’s email campaigns or Facebook posts about a new campaign that they’d like affiliates to participate in. The last step is commission reconciliation, which required the customer to manually track the commissions, account for outside bonuses, refunds, failed payments, usually waiting 60 days or more, download the PayPal Mass Payout report, then uploading it to PayPal, which then sends the payments to the affiliates. We are targeting this segment of the customer journey in this project and looking to automate it for customers.

1

Affiliate registration

• Create a landing page aka “Affiliate signup page”
• Qualify affiliates based on audience or register them manually

2

Set up commissions

• Set up share links and coupons for affiliates to use

3

Incentive affiliates

• Announce via email campaigns or Facebook posts about a new campaign or bounty program

4

Commission reconciliation

• Track and edit commissions
• Account for errors or outside bonuses and prizes
• Usually wait at least 60 days b/c of potential refunds
• Upload PayPal Mass Payout report to PayPal
• PayPal will send automatic payments to affiliates

Even though this was a comprehensive journey mapping of the end to end workflow for a customer running an affiliate program, and may seem like unnecessary since for this project we are only looking at payouts, but this exercise was helpful for me to quickly get ramped up on affiliates, share the knowledge with key stakeholders, gain empathy, and provided a foundation as we looked into other affiliate projects, which was also done during the same time we were working on payouts.

Front of the whiteboarding exercise on the current affiliate workflow.

Back of the whiteboarding exercise on the current affiliate workflow.

💡 Problem

1. Current process is manual and labor intensive: Creators had no way to efficiently manage payouts. Current process required them to double check the affiliate email and transaction amount and then submit a mass payout excel sheet via PayPal.

Example report customers download today, then cross-check against transactions manually.

2. Customers prefer to have both Manual and Automatic payouts, to have more control over how they like to pay their affiliates the commission.

Screenshot of a customer walking through a competitor’s payout feature

3. A large number of Kajabi heroes pay out their affiliates on a designated date of the month.

Insights from a customer thread on Facebook on how they keep track of affiliate payments.

🎨 Designs

From the problem definitions, I landed on the core product components for the affiliate payouts flow:

1. A system that supports both Manual and Automatic payout
2. A payout flow for Manual that allows customer to cross-check transactions
3. Setting a schedule for Manual
4. Setting a schedule for Automatic (set payout date on a specific day of the month)
5. View and export transactions and payouts

I started by designing for both manual and automatic payouts together, so the system could account for both from the start, then sequenced implementation with manual shipping first and automatic following shortly after.

For initial designs I used AI prototyping tools to do some generative brainstorming. I prompted it with the key product components and had it generate an initial affiliate payout flow. I find AI really useful for brainstorming because it gives me a quick baseline of what the product could look like. Then I can take those initial ideas, evaluate them, get feedback from stakeholders, and then dive into each component to refine the experience further.

For example, in this Process Affiliate Payouts modal, customers can have hundreds of payouts to process, so a small table within a modal isn’t going to work for that volume of information. It’s also not ideal for cross-checking transactions, because the modal can easily be closed and there’s limited space to review everything side by side.

AI generated output of initial explorations based on key product components.

I explored a few different ways to present automatic and manual payouts.

Final design of the Transactions dashboard offering quick entry points to both Automatic and Manual payout.

🎨 Key decisions

“Manual payment due” becomes “Manual payout due” I initially advocated for keeping the word ‘Payments,’ because from my perspective, the customer is making a payment to their affiliates, so I felt the term mapped more directly to the action. But after several discussions with the PMs, we ultimately decided that ‘Payments’ could be confusing because it introduces another piece of terminology for customers to learn. By using ‘Payouts’ consistently throughout the experience, we were able to create a more streamlined mental model and align it with the Payouts tab below, where customers can see a history of payouts made through either automatic or manual payout methods.

🚀 Outcome

• Creators enabled 1,000 manual payouts and 466 automatic payouts.
• Creators paid out $400K in commissions,
• The total Gross Merchandise Value (GMV) reached $4.3 million, indicating strong performance and engagement.
• We also observed a 9% growth month-over-month from September to October, highlighting the positive trend in our affiliate activities.

💭 Reflection

This was one of my favorite 0-to-1 projects. The lack of an existing system gave me a lot of creative freedom, but it’s also what made it one of the more complex builds I’ve worked on. We surfaced a lot of edge cases as we went, particularly around automatic payouts, for example, PayPal requiring re-authentication even when a customer already had it connected, which meant some real last-minute scrambling to redesign around it. If I were starting this today, I think the real opportunity is using AI to map the system’s edge cases upfront, rather than discovering them mid-build.

💡 Problem

1. Split across two products

   The onboarding experience needed to hand off seamlessly to Stripe to complete business and identity verification (spanning two separate products), not one flow fully within our control.

2. Lack of guidance & help

   Stripe gave barely any education about which business type/structure to choose from. If customers put in the wrong business information that doesn’t match IRS or local corporate records, Stripe’s automated verification system will reject them and will trigger a verification hold on their payouts. CX receive around 9 requests per month to remedy crucial portions of onboarding on behalf of the customer.

3. Fragmented experience

   Current onboarding was fragmented and non-linear. Customers connected Stripe first, got bounced back to Kajabi with “incomplete setup” errors, and often never finished — thinking they were done.

🎨 Key decisions

1. Swapped the order: business info + contextual help on business structure + T&Cs before Stripe connect, instead of Stripe first.
2. Bundled similar actions together to reduce steps and cognitive load.
3. Made Stripe signup the last step, so users arrive with everything ready.

🚀 Outcome

• 58.31% penetration, surpassing Stripe by 21%
• $1B+ creator volume processed
• 11% lift in KP adoption from the redesign, fewer CX tickets on onboarding
• New customers now default into KP with fewer objections; live across CA, GB, AU, UAE, EU

💭 Reflection

If I could do it again, I’d scope the first 30 days post-onboarding into the project from the start, including the monthly payout banner we designed but didn’t ship due to timing, since communication about when money arrives is just as important. Onboarding isn’t a funnel with an exit; it’s the first chapter of the product relationship.

💡 Problem

Current Payments dashboard was lacking in terms of features. Engagement with the dashboard was low. Customers have reached out requesting additional functionality such as Total Sales, best-selling Products & Offers, and LTV for a customer. They were also confused about how to navigate to the Transactions, Subscriptions, and the Multi-payments page from the current design. With our redesign, we want to improve the usefulness of the payments dashboard, increase engagement, drive return visits to the dashboard, and improve CSAT.

🎨 Key decisions

1. Use a modular redesign approach with swappable data viz components that allow for a phased rollout instead of an extensive redesign.
2. Display relevant, useful, and actionable data such as customer lifetime value and failed payments.
3. Provide a clear navigation for users to easily access different types of payments.

🎙️ Research

From 10/16 to 11/10/23, I ran moderated interviews with 14 experienced creators ($60K+ revenue, mix of Kajabi Payments and Stripe users) evaluating the new prototype against the current design. My PM also sat in and observed the interviews. The interview sessions also turned out to be a great co-creation experience with our customers, walking them through the prototype let us pressure-test whether the design mapped to their expectations in real time, and their reactions fed directly into the next iteration of the work.

🚀 Key findings

💭 Reflection

Doing this research gave us conviction: all 14 participants preferred the new design, so every decision after that was backed by evidence instead of opinion. I also presented the findings to leadership, which turned the research into shared design visibility for the org, also builds credibility for the team.