The radio frequency identification checkout system (RFID) at Berkeley Public Library needs more money. Yes, even before the system has been paid, some of the equipment is so deteriorated that it needs replacement. But the principal reason is that the original vendor, Checkpoint Systems, has abandoned the RFID maintenance market and sold its client base to another company. The library wants to spend more money for a maintenance contract and equipment from the new vendor. The sticking point is that the new vendor, 3M, does not comply with the Nuclear Free Berkeley Act, and so a waiver is needed for the library to proceed.
Many people might wonder why the library would want to continue with RFID at this juncture. Originally purchased as labor-saving technology, it’s proved to be a net consumer of labor hours. That is, it requires more labor hours to operate than it saves. This was especially true during the installation phase, when the soft costs—primarily labor hours appropriated from the library’s operating budget to install the system—amounted to more than the original price tag. Now, nearly four years into the experiment, the library has come up with some savings. It’s advancing the notion that RFID might have displaced labor hours equivalent to the work of 1.44 employees—an insignificant amount in view of the system’s costs.
In June of 2007, the American Library Association’s annual convention devoted some of its program to a consideration of RFID. The consensus was that RFID systems produced no positive return on investment. So reported the current library director in oral comments to the board at its July meeting. And, in fact, relatively few public libraries use RFID—between 1 percent and 5 percent. This is because most institutions have realized that RFID is expensive to purchase, expensive to install, and expensive to maintain. Santa Clara County Library studied RFID extensively—prior to adoption!—and rejected it as unlikely to produce any net savings. By contrast, Berkeley Public Library was early to adopt and late to catch on.
Going further, Oakland Public Library tested RFID prior to taking the plunge. It was tested at one of Oakland’s branches, tested entirely at the vendor’s expense, and removed when it failed to perform. In September of 2005, in the East Bay Daily News, Jackie Griffin, then library director, pronounced the entire RFID enterprise at Berkeley as a “…test…a beta version….”, in a quaint attempt to defuse criticism. Well, it’s been a very expensive test and the results indicate failure.
And so BPL’s rationale for pressing forward with RFID boils down to two arguments: (1) that, if RFID is abandoned, the money spent on it to date will be “wasted,” and (2) that it will be too expensive to transition to an optical/barcode system.
1. The first argument is based on a fundamental misconception. It’s a common one, where failed projects are driven forward based on the magnitude of their already expended (sunk) costs. The past investment in most cases is unrecoverable (e.g., there’s no market for used library RFID equipment). Hence, sunk costs are not relevant for judging whether the project should be continued. What is relevant are the expected future savings. But what will they be? The most likely prospect is that the current RFID system will have to be replaced in the not-so-distant future. The predictable hand wringing over the “waste” of public assets neglects the principles of basic finance. Prudent managers do not re-invest in poor-earning assets; they get rid of them. The opposite thinking leads to the kind of Bridge-to-Nowhere arguments where, if half the bridge has been built, the other half needs to be completed to avoid “wasting” the sums already spent. Folk wisdom puts it more succinctly: no throwing good money after bad.
2. Anyone who’s been to the grocery store recently has observed the speed and efficiency of optical/barcode checkout. It’s a robust, mature technology which costs a fraction of RFID. Libraries such as San Jose Public have well-attested rates of 80 percent for patron self-checkout with optical/barcode systems, higher than the highest rate for self-checkout ever claimed by BPL. So, why can’t we have it in Berkeley? The library is claiming that it will be too expensive to transition to optical/barcode. (As opposed to the continuing expense of RFID?) But their estimate of the costs for a new system, hardware and software, is just too steep for credibility. More realistic numbers might come from optical/barcode vendors. Moreover, BPL does not factor in any savings from such a transition. No savings for improved optical/barcode self-checkout. No savings for the expenses forgone by eliminating RFID—replacement of RFID equipment, maintenance contracts, RFID tags, etc.—projected to escalate in future years. A professional librarian’s independent estimate of annual savings from abandoning RFID in the periodicals section of the library, alone, is $10,000! Projected over the total items in the library’s collection, the savings will clearly be far greater.
If the library persists in its pro-RFID bias, when will the library user (citizen and taxpayer) get a break? Are we stuck with the RFID system largely because it didn’t work out? Why not do it right this time and invest in something that has a proven track record? In 2004, the City Council found outside financing for the RFID system. It could do this now for optical/barcode, a genuine improvement for the Public Library.
If we allow for palliative, time-wasting measures to keep RFID in place now, what happens? Who wants to bet that the library won’t be coming back to the taxpayer, near term, for a brand-new RFID system? They’ll certainly have to before 2011 when most of the library board’s RFID zealots will be termed out. Any takers?
Jim Fisher is a member of SuperBOLD (Berkeleyans Organized for Library Defense).