WORK
Equifax I-9 HQ Pricing Redesign
Equifax Workforce Solutions · Self-Service eCommerce · 2024–2025
Who hires exactly 10 people every month?
Nobody does. But that was the smallest plan on I-9 HQ and Forms HQ, a self-service line of business forecast at roughly $2 million for 2024. I was the revenue owner for the line and led its lifecycle motion. I did not own the enterprise billing infrastructure or the teams whose systems had to change. The pricing ran on monthly tiers based on employers’ forecasted hiring: 10, 20, 40, 60, 80, 125, 200 new hires a month, reset every month. There were fourteen SKUs in total across I-9 HQ and Forms HQ.
I don’t know why the tiers were built that way, or what the people who designed them were solving for. What I do know is what I watched happen once I was close enough to individual accounts to see it clearly: customers were being asked to forecast an inherently uneven behavior and buy capacity around that forecast. Use less than your subscription tier and you paid for I-9 verifications you never made. Use more than your monthly allotment and you hit an overage charge for each additional I-9 verification. I used to describe the experience to leadership the way cellular carriers used to handle data plans: an overage fee creates revenue, but it doesn’t create customer value.
I inherited this product with basically no budget and no roadmap headcount, inside an organization built for enterprise accounts rather than self-service SMB. Every month, I reconciled Salesforce, a cancellations database, the platform database, and Oracle billing by hand because there was no single operating view of the business. That reconciliation determined practical decisions: who should be billed, which payments had failed, which cancelled accounts needed to move through disposition or collections, and whether trials converted cleanly into paid subscriptions. I used the same account-level view to understand cohort retention, churn, I-9 Anywhere volume and plan mix, identify accounts that warranted lifecycle or commercial intervention, and report the state of the business to senior leadership. From when I took over the line in Q4 2023 through my departure in Q1 2025, monthly churn, including both voluntary and involuntary churn, moved from roughly 5% to 4%. During that period, the dunning and collections work I built recovered about $250,000 in one year. At Equifax scale, that recovery was small; relative to this self-service line, it was meaningful. The account-level view also surfaced self-service customers whose needs had outgrown the motion and who ultimately moved into enterprise contracts. It was hands-on work because the operating layer did not otherwise exist. When a change required systems I did not own, it had to compete for capacity on other teams’ roadmaps. That proximity to individual account behavior is what made the pricing mismatch visible.
The system underneath the slider
I didn’t know how the pricing was actually built when I started. It sat on top of something called a Charge Offer, an underlying billing construct that defines how usage grants get created and reset. I read what internal documentation existed, then sat down with the person on that team who understood the billing infrastructure, and worked through what was and wasn’t possible to change. One thing that opened the door: the reset period on a grant wasn’t fixed. It could reset quarterly, biannually, or annually instead of just monthly.
I wanted to go further and remove the punitive overage structure entirely. I couldn’t get that moved through the organization; it touched too much billing infrastructure for the scope and authority I had. So the annual-grant idea became the workable intervention: not the ideal end state, but a way to give customers more room for lumpy hiring without requiring the organization to rebuild the billing system first.
A better answer, within the system we had
In Q2 2024, I proposed annual plans at a 20% discount, plus two new smaller monthly tiers: 1 hire and 5 hires, sized for the smallest self-service accounts. That took the SKU count from 14 to 36. The pricing structure was more complicated than volume alone. Starter and Standard carried different capabilities: all plans could use I-9 Anywhere In-Office, while Standard included E-Verify integration, which enabled I-9 Anywhere Virtual. At the same time, I worked with the senior leaders responsible for web and digital strategy to simplify the buying experience around that complexity: clearer language about what customers were selecting, fewer self-service package choices, and a single Enterprise path where larger or more complex accounts could move into a commercial conversation instead of choosing between multiple contract-oriented cards.
The pricing work sat inside a broader optimization program rather than a single page change. We used Optimizely to test parts of the selection experience, including CTA language and where the hiring-volume slider should default. In one test, we changed the default from the legacy 10-hire tier to the new 1-hire tier. The 10-hire default won, so we kept it. Creating a smaller plan did not mean we should assume every customer belonged there or steer the experience toward it. We could not cleanly test every question we wanted to test, including monthly versus annual selection, because of implementation constraints in the experience. Conversion and trial-to-paid performance improved during the broader period of pricing and experience work, but several things changed together, so I do not attribute that movement to the pricing redesign alone. In the limited post-launch window I saw, overage charges also declined. When an account still hit an overage, that became a useful signal that the account likely needed to be right-sized or moved into a contract conversation.
The pricing had more dimensions than the slider showed
The first redesign improved fit, but it also made the structure underneath more visible. Price reflected more than hiring volume. Customers were choosing among packages with different capabilities, estimating how many people they expected to hire, choosing a billing period, and using I-9 Anywhere as a separately metered service. Starter and Standard were not interchangeable products. Standard’s E-Verify integration enabled the virtual I-9 Anywhere workflow; Starter was limited to the in-office workflow.
That changed the commercial question for me. Instead of continuing to add combinations of package, capacity and billing period, I started separating what customers were actually paying for. Platform access was one thing. Completing an I-9 in-house was another. Having Equifax perform the verification through I-9 Anywhere was a third, higher-value service. I also argued that the price gap between do-it-yourself completion and I-9 Anywhere should narrow while preserving a meaningful premium for the done-for-you service.
The final public architecture reflects that separation: platform access, per-form pricing for in-house completion, and a separate per-form price for I-9 Anywhere. The public page shown below lists $8.50 per month for Starter platform access, $7.95 for each in-house I-9, and $84.95 for I-9 Anywhere. I no longer have the historical package-by-volume matrix or customer usage distribution needed to make a defensible aggregate ARPU, margin, or retention claim, so this asset does not make one.
What the first redesign made harder to ignore
The 36-SKU redesign improved fit, but it also exposed the cost of continuing to solve the problem by adding combinations. The business was layering package entitlements, hiring-volume tiers, billing periods, and separately metered services onto the same buying decision. The intermediate redesign was a workable answer inside the billing system we had. It also made the underlying pricing architecture harder to defend.
I-9 Anywhere became an important clue. It was already priced separately from the subscription tiers and metered per use. Working through the first redesign changed how I interpreted that fact. Customers were not buying one homogeneous unit called ‘hiring volume.’ They were paying for access to a set of capabilities and, in some cases, for verification work Equifax performed on their behalf. That made forecasted monthly hiring feel increasingly arbitrary as the organizing unit for self-service pricing.
In Q1 2025, that became the final pricing architecture I proposed before I left: remove the hiring-volume tiers, separate platform access from completed verification work, and price verification work by type, with one rate for do-it-yourself I-9 completion and another for the done-for-you I-9 Anywhere service. This was not simplification for its own sake. It separated access, usage and service level instead of forcing them through a forecasted-capacity tier.
The model I left behind
I designed the architecture, worked through its pricing implications, brought it to senior leadership, and got approval to move it forward. As one of my last acts before leaving, I translated the approved model into roughly a dozen implementation tickets spanning the functions required to change pricing, billing, product, web, customer operations, and the surrounding systems, then worked with those owners to get the change into motion. I left before implementation was complete. The model shipped afterward, so I do not claim the downstream execution or post-launch results. My claim is limited to the diagnosis, the pricing architecture, the leadership approval, and the cross-functional work I completed before departure.
The first redesign was not the final answer. It gave smaller customers more appropriate options inside the structure we had, while exposing how many different commercial decisions had become entangled in that structure. The next redesign separated those decisions: platform access, in-house verification usage, and the higher-value done-for-you service. The final public pricing page is the receipt for where that reasoning landed.