1. A nonprofit runs an enterprise-grade stack on nonprofit margins
Road Scholar, a nonprofit experiential learning organization enrolling over 100,000 people annually, has spent roughly ten years building on top of Optimizely's platform. According to a MarTech Zone interview with Mark Fagiano, Associate Vice President of eCommerce at Road Scholar, the organization moved off a homegrown stack to Optimizely's CMS about a decade ago, then adopted additional products as Optimizely acquired and integrated them: content marketing, customer data platform, experimentation, and the Agent Platform (formerly Optimizely Opal).
The organization's channel burden is real. Road Scholar's audience, averaging 50+ years old, still wants physical catalogs but also browses on phones, converts on desktops, and calls a contact center with questions. Marketing covers Google, display, retargeting, SEO, email, connected television, direct mail, and oversized printed guides. Donations arrive at enrollment rather than through a separate fundraising funnel, meaning acquisition, retention, and mission work (scholarships for caregivers, awards for educators) share a constrained budget, according to the same source.
What makes this case worth studying is not the vendor selection. It is the operational sequencing.
"We’ve doubled our conversion rate through mobile over just the last two years."
2. Consolidation as an operational act, not a buying spree
Road Scholar did not purchase the full Optimizely suite at once. As reported by MarTech Zone, the team started with the CMS, then expanded into content operations, data, and experimentation as Optimizely absorbed those capabilities through acquisitions (Episerver acquired Optimizely in 2020, rebranded in 2021, and continued adding content operations, a CDP, and B2B commerce). Road Scholar adopted products on offer rather than stitching together separate vendors.
This pattern matters for enterprise marketing operations teams evaluating their own stack architecture. When your MarTech problem is a diagnostic problem, the instinct is to buy another tool. Road Scholar's trajectory suggests a different discipline: adopt within the platform you already have, test the gaps, and only go outside when the platform genuinely cannot cover the job.
Fagiano's attribution model illustrates this. Attribution does not live inside Optimizely alone. Demand hits the CRM, then a third-party model assigns value across touches and devices. According to Fagiano, last-click attribution "starves the catalog and the first ad, and overfunds the form." The platform's role is to keep the site, tests, profiles, and content calendar in one system so the attribution model is not matching dirty data. That is a data management discipline, not a platform feature.
3. Mobile conversion and the agent layer: two concrete results
The most specific performance claim in the interview is mobile conversion. According to Fagiano, Road Scholar doubled its mobile conversion rate over two years. The team discovered through experimentation that checkout was not built for thumb navigation, then rebuilt the flow. Traffic still arrives mostly on phones, and many users finish on desktop. But the phone now captures leads and no longer blocks the users who stay.
The contact center shift is equally telling. Over about ten years, most reservations moved from the contact center to the website, per Fagiano. The contact center now spends time on questions only a human can answer.
On AI agents, Fagiano reported that Road Scholar pushed agents through marketing and communications over 18 months until usage became daily, then extended them to other teams. The organization built agents so a lean staff can cover more of the mission. They defined four participant personas and use synthetic versions of those personas inside the Agent Platform to pressure-test copy and landing pages before a human reviews them. This connects to our earlier analysis of how agent hubs are becoming the control plane for revenue operations.
"Opal is in everyday use. We built a lot of agents around Opal, so we can make sure that we’re getting the most efficiency out of the people that we do have."
4. What enterprise ops teams should take from this
Consider auditing your existing platform's unused capabilities before adding a new vendor. Road Scholar's approach was to adopt what Optimizely offered as the suite grew. Many enterprise teams on Oracle Eloqua or Adobe Marketo have purchased modules they have never activated. A platform maturity assessment that maps current feature usage against available capabilities will often reveal that the next win is adoption, not procurement.
We recommend separating the attribution question from the platform question. Road Scholar keeps attribution outside Optimizely, in a third-party model connected through the CRM. The platform's job is to keep data clean enough for the model to work. If your organization is running multi-touch campaigns across paid, organic, email, and offline channels, the attribution model should sit where it can see all of them, which is rarely inside one marketing platform.
Consider building synthetic persona testing into your content workflow. Road Scholar's use of AI-generated persona surrogates to evaluate copy before human review is a practical application that does not require a full marketing AI transformation. Start with two or three well-defined personas and test them against a single campaign type. Expand only after the process proves faster than your current review cycle.
We recommend treating mobile conversion as a standing experiment, not a one-time redesign. Road Scholar's two-year doubling of mobile conversion came from ongoing testing, not a single project. Teams running always-on campaigns should instrument mobile checkout and form flows for continuous experimentation, with the goal of finding and removing friction as user behavior shifts.
The broader lesson is structural. Platform consolidation reduces the number of identity graphs, login screens, and data-matching problems your team manages. But consolidation without operational discipline produces the same dysfunction in fewer systems. Road Scholar's results came from a decade of incremental adoption, continuous testing, and a clear understanding of where the platform ends and where external models (attribution, CRM) begin.
