The Missing Asterisk: How OverDrive Built a Business Model on Unverified Claims

The Panorama Project's flagship study measured what happened when OverDrive ran the "Big Library Read" program in April 2018. One book was simultaneously featured across 14,700 library branches for two weeks. The result: ebook sales jumped 720% and print sales rose 38% above pre-campaign levels.

Sounds like proof that library lending drives retail sales. But here's the problem:

What This Study Actually Measured

OverDrive ran a massive, coordinated promotional campaign through its own platform across 14,700 branches and measured the sales impact. That\'s not typical library lending. That\'s a marketing campaign.

It's like Coca-Cola running a $10 million Super Bowl ad, measuring the sales bump, and then claiming "grocery stores see a boost." Yes, because they ran a national advertising campaign.

The study proves: National promotional campaigns work. It does not prove: Typical library circulation drives retail sales.

Why This Matters

Potash cites this number everywhere. "Library lending drives retail sales" becomes the justification for the bookstore connection pilot. If library patrons read books for free from libraries, and free reading drives retail purchases, then connecting those patrons to bookstores is good for everyone.

But the evidence for that premise comes from OverDrive's own marketing campaign, not from typical library borrowing patterns.

The Panorama Project's 2020 consumer survey found that 38.31% of respondents had bought a book online that they first found in a library within the last 12 months.

This number gets cited a lot. But it has three critical limitations:

Problem 1: Sample Bias

The survey recruited over 4,300 respondents via Qualtrics panel, 75% of whom already had library cards. This is not a random sample of book buyers. People who use libraries are more engaged readers and more likely to buy books anyway. The survey doesn't measure whether the library exposure was the deciding factor or whether those people would have bought the book without ever visiting a library.

Problem 2: No Causation

Finding "a book in the library, then bought it later" doesn't prove the library caused the purchase. It could be:

• The person was interested in the book anyway and used the library to preview it before buying
• The person saw a positive library review and that made them interested in buying
• The book was on the bestseller list and the library carried it; both sources are correlated but independent
• The person read the library copy, enjoyed it, and bought the next book by that author (not displacement, but a different mechanism)

Correlation is not causation.

Problem 3: No Control Group

There is no measurement of what percentage of book buyers don\'t use libraries. There\'s no control group of people who never found books in libraries but still bought books. Without that comparison, 38.31% is meaningless.

Potash wrote, in official press releases and testimony, that "retail sales see a boost*" from library lending.

That asterisk implies there's a citation, a study, a data source backing up the claim.

But the footnote was never published. The most likely source is OverDrive's own Panorama Project, which:

• Was founded by OverDrive
• Was funded by OverDrive
• Was designed to prove a specific outcome
• Measured OverDrive's own promotional campaigns
• Found results that independent research has contradicted

He can\'t cite it because citing his own company\'s funded research in a press release about his own company's product would make the circularity obvious.

But he can't NOT make the claim, because the entire pitch to publishers, and the entire justification for the bookstore pipeline, depends on it being true.

The missing asterisk is the tell.