The Worst Thing You Can Say to Me: It Can't Be Done.

tl;dr: I suggested Baker & Taylor build an ILS in 2018. Years later, I started trying to build one myself.

On the last day of the 2018 Baker & Taylor and Follett sales conference in Las Vegas, I suggested Baker & Taylor build its own integrated library system (ILS).

They had most of what was needed, it almost seemed like a no brainer.

Everyone in the room looked at me like I’d lost my mind. To be fair, most were either hung over or still drunk.

And to also be fair, they had reasons. But I didn’t learn this until YEARS after.

When Follett bought Baker & Taylor in 2016, it inherited years of old systems and workarounds. TitleSource and TitleSource360 had layers of systems, workflows, and decisions piled up. I can’t remember the year, but the person who lead TS360 quit during one of the major upgrades which caused issues. MAJOR issues.

I don’t think Follett understood how tangled that technology was when it bought the company. Starting over would cost money in a business with thin margins. It didn’t make sense at the time; and it still doesn’t. But this wasn’t Follett’s first rodeo with BT, they had done this song and dance a few times before. They had to have known what they were getting into with that tech debt.

So B&T put its attention elsewhere. One project was BTCat, building on its book purchasing and processing service, in-house cataloging team, and the spreadsheets and tools already holding work together.

That seemed reasonable. Then COVID hit.

Baker & Taylor was too big to qualify for the COVID relief loans that kept smaller distributors alive. The company still had to pay for warehouse space while nobody knew what came next. (We need to talk about the warehouses in another post; jaw-dropping details from the lawsuit. And a couple “I told you so’s”.)

They furloughed us in alternating weeks: one week on, one week off. Honestly? I loved that schedule.

My wife had been shipped to NYC for the COVID response and I instantly became a single-mom to a toddler (I don’t deserve all the credit; his dad and step-mom were amazing during this time, too.) The schedule worked out perfectly.

When we came back full time, it was chaos. Libraries were trying to match inventory with invoices. The company was dealing with returns, new stock, warehouse space, and training new workers. Aman Kochar later told Publishers Weekly, “We were coming out of a brutal period.” Fuck yeah we were.

And it only got worse.

I left at the end of the first quarter of 2021.

At this point in my career, I’d spent years as a librarian and administrator, then worked with libraries through B&T’s collectionHQ tool. It didn’t matter.

What sucks the most… the people who wouldn’t listen, are still in the industry today. Similar job title; different vendor. Still grifting.

On November 4, 2021, Kochar completed his acquisition of Baker & Taylor, ending the company’s five years under Follett, according to Publishers Weekly.

In December 2023, the magazine reported Kochar’s estimate that half of B&T’s products, services, and solutions hadn’t existed two years earlier. This is what happens when you don’t have a solid business plan nor leaders who understand the industry they serve.

But he was from India, came from McGraw-Hill and knew technology - so he was “our guy.” (I’m not kidding. Don’t forget company was based in North Carolina… stereotypes did help in some cases.)

He was proud when talking about their diversity audits; which was basically running cHQ/ESP reports, mashing it with metadata, sorting, adding in some formulas and calling it a day.

Rebrand Axis360 into “Boundless.”

All of this shit had been around since 2000s… collectionHQ/ESP had been in the UK since 2005ish, sold in the US since 2013ish; BT had ebooks going back 2010s when they first partner with OverDrive.

They just kept putting lipstick over their services; rebrand after rebrand.

The patched-together systems where still underneath all the “new products.”

I studied the BT platform as a competitor in 2009 when it launched after OverDrive. I used collectionHQ as a customer in 2013, and then sold it as a vendor in 2018. I helped develop the algos for ESP in 2019. Librarians were hacking together diversity audits way before you could put a price on it.

Seriously insulting to all of the people, librarians, hell the entire industry to say he had products and services that hadn’t existed two years earlier. Fuck off with that bullshit.

So by 2025, I’d spent over a year learning about AI and everything in between. It turned into an obsession. Some might say unhealthy, but hell, here we are.

I wanted to understand the code; see how all of the pieces and parts worked together using APIs; crypto security for website apps. The more I learned, the easier things started to click into place until I finally said - fuck it, I’ll build it myself.

Before AI, I’d search Google, Reddit, and coding forums, then piece the answers together. I’m an elder millennial librarian… if there’s a will, there’s a way.

But now I could ask questions against those sources and follow the citations (which is terrible for someone with self-diagnosed ADHD. So many links to click.)

The hardest part was figuring out how to stop adding features and functionality.

Working with library technology for almost two decades, I knew exactly what I would want. And I wanted an all-in-one, accessibility first, web app that could replace all of the software subscriptions. And it had to be mobile friendly. Hard stop.

I just kept adding as the pieces fit together until I ended up with a full-blow integrated library operating system.

When I got stuck, I could ask the LLM for a simpler explanation or a diagram. This was incredibly useful when mapping out database relationships and understanding workflows. I would have it walk me through resolving bugs.

It wasn’t smooth. I moved two steps forward and one back plenty of times. I now have a working ILS with circulation, cataloging, patrons, acquisitions, search, reporting, and more. I wouldn’t tell a library to migrate to it tomorrow. I’d like to pay for a professional pentest, but one costs a couple thousand dollars. And mama ain’t got funds for that.

So the code and data models are open to inspect or customize. The system runs in containers, so its parts can be changed and rebuilt as the technology changes. You can completely trash the code and revert right back to where you were.

It’s pretty badass.

If I had a security certificate for it, I’d be pitching it everywhere.

Until then, message me for a link to the repo. If I can, you can, too.

Here’s a link to the live site - l30.org.