Monday, November 16th, 2009

Local Book Search: A new way forward for bookselling

We’ve just released a slew of new features (see overview). We hope members like them all, but Local Book Search is the most important. I won’t mince words: it is intended to change bookselling forever.

The Idea. It’s a simple idea: a location-based search for books. After all, you can ask Google for pizza where you live. But you can’t do it with books, until now.

Just Show Me the Feature! Here are some sample links:

Imagine if all the bookstores and libraries in central London were searchable.

The Vision. More and more people look to the web, even for local information. I can’t remember the last time I checked a newspaper for movie times or local music. Bookstores, however, have largely stayed off the “local web.” Those that expose their holdings do it inconveniently. Many don’t at all, either because their systems are old, or on principle. And nobody has ever tried to tie it all together, like Local Book Search attempts to do.

Physical bookstores often complain about online prices, but a lot of the problem is just convenience. If I lived in San Fransisco and wanted a copy of Steinbeck’s The Moon is Down, I probably wouldn’t bother to call up all my local bookstores; Amazon is quick and easy, and will surely have it. Local Book Search find a half dozen copies, new and used, in just seconds.

At a minimum, this is a cool test. Maybe bookstores will prefer to keep their holdings information to themselves and shut us out. A bookstore near me won’t participate because they think the information is “valuable” and, if it were online, “people would know what we don’t have.” I think that sort of thinking is the road to a world without bookstores.

Online bookselling is here to stay. It’s a wonderful thing. Same-day online delivery, as recently offered by Amazon, is something new. But physical book stores are also very important. We hope local book search can be one of the things they need to survive and prosper.

Who are you searching? The list currently stands as follows:

How are you doing it? We’re doing it in a mix of ways. A number of bookstores, including Powells, Porter Square and City Lights, send us their files. We have special permission from Abebooks too, so we pre-parse the Abebooks sellers.* (Members have helped us connect them to LibraryThing.) Indigo/Chapters gave us their blessing, and a convenient way to do it.

The rest involve just-in-time searches. Every system is different, so we’ve written a special connector for each, but the usual pattern is some sort of multiple-ISBN fetch. In every case, we’ve obeyed a site’s “robots.txt” file–going only where we’re allowed, and hitting a website no faster than once per second. That’s also why we’ve stayed off some biggies, like Barnes and Noble–their store holdings pages are prohibited by robots.txt.

Why aren’t there more indies? We introduced basic bookstore integration back in 2006 (!), taking only indies. It never really took off, because we couldn’t convince IndieBound to offer access to the files, and, I think, asking booksellers to upload their file to IndieBound and then LibraryThing was a hard sell.

Since then, we’ve pushed hard to get indies involves, with no success. We even lost our first participant, when Ann Arbor’s Shaman Drum went out of business. Those who follow our Twitter stream will remember a lot of hand-wringing over the topic. (In fairness, IndieBound has expressed interest, but they are still working on their platform.) So, we went and did it outselves, with whom we could.

The “connectors” all take about the same amount of time, whether we pick up 400 stores in Borders, or a single store in the Strand. So, by time, we’ve spent most of it integrating with indies. We have a decent list of other indies to add, when we get the chance.

The Future. Obviously, we know this is bigger than LibraryThing. If you have an idea about where it should go, chances are we’ve had it too. We’re working on it.

We are, incidentally, aware that we’re weak on the library front. We have 40 libraries now and, if all LibraryThing for Libraries sites agreed, we could get up to 170 systems (more than 1600 buildings). Unfortunately, you can’t really screen-scrape most library systems; they tend to break under the strain. Nobody really has the data necessary, except perhaps for OCLC.

So, OCLC, how about it? Our peanut butter. Your chocolate. Don’t libraries deserve to be part of this? Can we let bygones be bygones?

What we need? Bookstores who want to participate should send me an email (tim@librarything.com) or a Tweet (@LibraryThingTim).

Bookstores concerned about LibraryThing servers hitting them too hard should also write us. We’re keeping close tabs on how often we scrape sites. (We also cache results for 24 hours.) In many cases, we could do it more efficiently, if we were allowed to access search pages, not just product pages. And we can try fewer ISBNs. Or, if you are able, we can parse your files and never hit your site.

All told, we think the traffic is good for you—it will get you sales. But we are conscious of limited technology budgets, and will work with you to increase the benefit, and decrease the potential cost.

Kudos. Thanks are owed to:

  • LibraryThing members who are helping us connect Abebooks to LibraryThing.
  • Abebooks. In theory, local book search could cut them out of the equation, as people walked into a store, instead of buying online. But they saw it as a valuable service to their members, and an overall win. Abebooks deserves applause for taking a risk.
  • Indigo/Chapters. Every time I deal with them, I marvel at how easy they make things. No meetings, just agreement to test it out. Indigo, like Abe, is Canadian. Is there something in the water?
  • Powell’s, Porter Square Book and (tiny) City Lights Books in Sylva, NC deserve thanks for being with us from the first. Because their method–sending us files–is so much easier for us, we’re planning to create some special features for them.
  • Programemrs Chris and Mike, and non-programmers Dan and Stephanie. The “Get this Book” and “Local Book Search” project took over two months, and involved huge amounts of new code, and also data entry. As Chris remarked, for most companies, “this would be a site.” It is an awesome achievement.

Come talk about it. Come talk about all this on Talk.


*Abebooks.com sellers already total over 1 million books, and we haven’t caught up with all the Abe-to-LibraryThing work members have done. We think they’ll end up around 1.5 million books total.

Labels: abebooks, booksellers, bookstore integration, bookstores, local book search, used bookstores

6 Comments:

  1. Anonymous says:

    It's like, how much more awesome could this be? And the answer is none. None more awesome.

    LibraryThing truly does go to 11.

    craigim

  2. Brian E. Raitz says:

    So how do we set up our local library? Basic search for Parkersburg & Wood County PL is
    http://catalog1.mln.lib.wv.us/search/i?031615461X&searchscope=26

    The other libraries in MLN are the same with different scope numbers.

  3. Katya says:

    "Nobody really has the data necessary, except perhaps for OCLC."

    And even then, they're missing a lot of public libraries.

  4. Anonymous says:

    Is there a way to set up our personal library, already on Library Thing, as an independent book seller?

  5. Anonymous says:

    I'd love to be able to offer books for sale that are already listed in library thing. How do I get to do that?

  6. Jim Morris says:

    Is this idea working? I occasionally want a book right now and can’t wait for Amazon. Longer term, everything will be delivered electronically; but, for older books, maybe there’s an opening for independent stores if they can get their inventory online.

    What were/are the biggest impediments to your vision?

Leave a Reply to Anonymous