Fact Check: If Almost Forty Percent of ABA Members Aren’t Actually Indie Bookstores, Can We Really Say There’s a Revival?
For the past several years it has been fashionable to write about the indie bookstore revival, but there’s a huge problem with many of the stories: they cite data which doesn’t stand up to scrutiny.
Starting in 2013, many bloggers and journalists have cited the membership count of the American Bookseller Association in support of the narrative that indie booksellers were going through a revival.
For example, one of the points Glenn Fleishman raised in his article in the Seattle Review of Books last week touched on this data point:
Now indie stores are seeing a resurgence — small, but noticeable, sustained, and continuous. The ABA says its membership has grown from 1,401 and 1,651 locations in 2009 every year through the current one, with 1,775 members and 2,311 stores even as Borders shut down and Barnes & Noble closed 120 outlets during the same period. (Not all indie stores are ABA members, though most are. This can result in apples-to-oranges comparisons among articles that rely on ABA numbers, industry figures, or government analyses.)
I was one of the first to report on the revival in 2013; at that time I saw the ABA membership stats as evidence which disproved the then-accepted belief that Amazon was killing indie bookstores left and right.
Today I would like to clue you in on a detail which a friend on Twitter pointed out a while ago: there’s a problem with the ABA data.
Sure, we’ve read many stories about new bookstores opening, including about a franchised chain of used bookstores. And yes, nineteen of the twenty new ABA members are bookstores.
But the fact remains that a significant number of ABA members simply are not bookstores.
While the ABA has positioned itself as representing indie booksellers, it will actually accept retailers of all type. In fact, its membership application form invites prospective members to identify as "non-bookstore", and the form welcomes all type of businesses, including online sellers, museums, colleges, and even "non-retail operations". (I am seriously tempted to join under that last option.)
As a result, the actual ABA membership rolls are more diverse than the group’s name would suggest. In Virginia, for example, the ABA lists 51 members. At least twenty of those members are not indie bookstores.
Four college bookstores in Virginia are ABA members, including two stores run by Follett. The membership roll also includes two online booksellers (one is just down the road from me) and an online retailer as well as the Marine Corps Association, three national parks, a toy store in Staunton, a museum, a community center, a holistic store, and five Hudson News airport concessions (actually around a dozen locations in five airports).
I’m sure all of these entities sell books, yes, but that still doesn’t make them indie bookstores any more than Walmart.
Remember, the phrase "indie bookstore" connveys the image of a small local bookstore which is run by the owner, and possibly looking something like this:
So tell me, with almost 40% of ABA members in VA not being indie bookstores, can we really say that there’s a revival?
Not from the ABA membership roll, no, but not all bookstores belong to the American Bookseller Association.
Just in my immediate geographic region, I can find four indie bookstores which don’t belong to the ABA. (What’s wrong with the ABA that I have more non-member indie bookstores within driving distance than ABA members? Is the membership really worth so little?)
If we could find a real source of statistics on bookstores, one which counted all indie bookstores and tracked the number of locations (one ABA member in Virginia has three bookstores) then perhaps we could answer that question.
But as it stands, there’s no way to know for sure.
images by robwalsh0, TravelNevada, leighklotz
RAD July 31, 2016 um 7:24 pm
"If Almost Forty Percent of ABA Members Aren’t Actually Indie Bookstores, Can We Really Say There’s a Revival?"
Sure, as long as the number of **actual** indie bookstores has been growing year after year. The data on the ABA web site is only the current snapshot, you need at least one year of historical data for comparison and ideally all of their historical data.
But you are absolutely right, the claims of a revival cannot be true without better data analysis. The analysis is not hard though. You can scrape the current web page for the dataset and have a baseline for doing the analysis moving forward. Some of the original data fields you show in the web form would be nice to have though. If you contact ABA and ask nicely maybe they will provide the historical data. Their contact page lists a guy who is the "Database Manager", sounds like your guy. Start a Github project, ask for help. ABA has data, a little legwork gives you the "source of statistics".
Be the source, Nate 😉
Nate Hoffelder July 31, 2016 um 10:03 pm
Maybe I should.
Do you have any suggestions on how to count the bookstores which aren’t ABA members?
RAD August 1, 2016 um 6:07 am
The full list of ABA members can be returned using a search with empty search fields:
This page says "Search fetched 2230 results".
If you do a View Source in your browser, you can see the table with the results, each table row has:
1. member name
2. member href (i.e. link)
3. member address (can be further parsed for a. city, b. state, and c. zipcode)
4. member phone_number
5. member description
If you scraped this page into a aba_members SQL table, you can do the following:
— full count of ABA members
SELECT count(*) FROM aba_members
— count of Indie only ABA members
SELECT count(*) AS indie_member_count
— exclude book chains https://en.wikipedia.org/wiki/List_of_bookstore_chains#United_States
WHERE name NOT LIKE 'Hudson%' — Hudson News is a chain of airport bookstores
AND name NOT LIKE 'Half Price Books%'
AND href NOT LIKE '%.bkstr.com' — Follet is a chain that operates many University bookstores
— exclude University book stores
AND href NOT LIKE '%.edu' — many Universities use .edu as a top level domain
AND name NOT LIKE '%University%'
AND name NOT LIKE '%College%'
AND name NOT LIKE '%Campus%'
— exclude other members determined to not be indie
AND name IS NOT 'Marine Corps Association'
The scraper can be written in Python and a library like Beautiful Soup. There are equivalents to Beautiful Soup in other popular languages like Node.js, Ruby, and Golang.
The scraped data can be exported into a CSV file to be loaded into SQLite or a SQLite database can be populated directly from the scraper utility.
Eventually, the "Indie Only" SQL statement has to be flushed out and vetted then applied to specific points in time. The web site has to be scraped periodically, either a specific dates or checking for daily changes using the last retrieved ETag and If-None-Match requests. A database approach for historical data (e.g. one database per year, one table per year, extra column with ETag/RetrievalDate).
If you can get the original ABA data then you use whatever structure they supply for the SQL statements.
Once you have the dataset, any number of useful questions can be asked and answered.