Brief Bios, Related Patrons, and Much Faster Search

Hi! A bunch of new things since my last update in late September. I’m flying to NYC right now after attending Cloudflare Connect 2025 in Las Vegas.

We’ve been focused on making the site more useful for discovering connections between patrons and institutions. Plus I met the Founder of Astro, one of the many tools used to make this website! So that was fun.

But back to the updates and donor data. Here’s what’s new.

🤖 Brief Patron Biographies

Our top patrons now include a brief biography. We’re using Google Gemini 2.5 Flash to help write these bios, pulling information from publicly available sources to create concise summaries of each patron’s background and philanthropic interests.

Here’s how it works: Gemini analyzes the patron’s name, their giving patterns across institutions, and publicly available information to generate a 4-7 sentence biography. Then we do a manual human review of many of those bios to help ensure accuracy before publishing.

The bios include:

• Key facts about the patron’s background
• Any fun facts, inventions, or major business history
• Their philanthropic interests and focus areas
• Gender-based pronouns for better readability

Check out examples on pages like Agnes Gund or Leonard Lauder to see what these look like. The bios add context that helps you understand who these patrons are beyond just their donation history.

Why this matters: These bios help researchers quickly understand a patron’s profile at the top of their profile page. We hope to pull more data from Wikipedia and other sources soon for our top donors.

🔗 Related Patrons

Almost every patron page now shows a Related Patrons section that helps you discover other donors with similar giving patterns. I’ve been wanting to launch this for a while and I think Mohamed finally figured it out with a good algorithm.

The goal is simple: help fundraisers or researchers discover other donors who support similar causes or institutions. If you’re researching philanthropy in contemporary art, you can find an entire network of related collectors and donors.

Here’s how we roughly determine related patrons:

• Shared institutions: Patrons who donate to the same museums and cultural organizations
• Giving levels: We weight connections based on the tier/membership level at each institution
• Recency: More recent contributions count more heavily
• Breadth: Patrons who overlap across multiple institutions rank higher

We also assign each related patron a title based on their most recent contribution. So instead of just seeing “John Smith,” you’ll see “John Smith – Metropolitan Museum Fellow” which gives you immediate context about their current giving level.

You can see this in action on pages like this donor profile in the new section called “RELATED PATRONS.” You have to scroll down a bit to find it.

This is still evolving, and we have to set a cron job to create the relations more frequently and for a lot of patrons who haven’t been indexed yet. But it’s already helping people find patterns in cultural philanthropy that weren’t visible before.

🔍 Search is Better and Faster

Remember that preliminary search feature we launched in September? I mentioned it needed work. Well, we rebuilt it from scratch.

The new search uses Cloudflare Workers KV to cache a searchable index of all patron names. Instead of hitting the database for every search query, we query a pre-built index that returns results in under 50 milliseconds.

Try it yourself at patronview.com/search. Type any patron name and you’ll get instant results as you type. The results are ranked by relevance and limited to the top 10 matches.

We also added a Search button to the main navigation menu and mobile menu, so it’s much easier to discover now.

Technical details for the nerds:

• Search index is generated nightly from the database
• Stored in KV as a JSON payload with patron metadata
• Client-side fuzzy matching for instant results
• No database queries during search – just a single KV read
• Falls back to database if KV cache is empty

The performance improvement is dramatic. Our previous search took 500-800ms per query and hit the database every time. Now it’s consistently under 100ms with zero database load.

🖼️ Custom Share Images

When you share a patron or institution page on social media or via text message, it now generates a custom share image automatically. These dynamic OG images include the patron’s name, contribution count, and a little bit of PatronView branding.

The images are simple and clean. They’re generated on-demand using Cloudflare Workers, so there’s no manual image creation needed for the tens of thousands of donor and fund pages.

Try sharing any patron page and check out the preview image. For example, David Geffen’s page generates a simple card showing his name and contribution stats.

⚡ Fund Pages Load 94% Faster

Building on the KV caching work from last month, we’ve now extended caching to fund pages. This reduces database reads by 94% for these pages.

Workers KV is Cloudflare’s key-value storage system. Think of it like a giant hash table distributed across Cloudflare’s global network. It’s optimized for reads – much faster and cheaper than querying a SQL database.

Instead of running complex queries with 10+ table joins every time someone visits a fund page, we now grab pre-computed data from KV in under 50ms. As our database continues to grow past 500,000 contributions toward millions of entries, this caching becomes even more critical for maintaining fast load times.

The result is that fund pages load consistently in 200-300ms instead of 800-1000ms.

🏛️ “Also Appears In” Section

Many individual patron pages now show an “Also Appears In” section listing other patron records where that person’s name appears. This helps you see the full scope of someone’s philanthropic portfolio.

This feature is still a work in progress. We’re working on cleaning up and associating fuzzy matches where donor names overlap across institutions. Sometimes “John Smith” and “John A. Smith” are the same person, sometimes they’re not. We’re building tools to help identify and merge these duplicates but it has been hard to figure out how to do well. This is a good stop-gap.

Other Updates

A few other small improvements:

Popular Programs on Institution pages – Institution pages now highlight the top programs and funds at each museum. This helps you quickly see where donors are focusing their support. Check out The Met’s page to see this in action.

Pagination SEO improvements – We fixed pagination to use proper noindex tags and canonical/prev/next links. This prevents Google from indexing thousands of duplicate paginated pages and helps our main pages rank better.

Mobile UX improvements – Fixed horizontal scrolling issues on mobile devices and added a close button and Home link to the mobile menu. The site should feel much more polished on phones now, but we still have to fix the Contributions Timeline table at the bottom.

See more/less toggle – For patrons with long contribution lists, we added expandable sections so you can see more without endless scrolling. Click “See more” to expand the full list. This helps for the top 2% of our patrons who have over 15 contributions in various categories.

What’s Next

We’re continuing to focus on data quality and expanding the database with more source files. We’re also testing some interesting new features and writing tools for summarizing giving patterns and generating insights about donor behavior.

I’d love to hear what you think about these updates. If you have feedback or feature requests, connect with me on LinkedIn or email me at nick@patronview.com.

Thanks for following along with these updates!

– Nick Gray