Following HarperCollins’ recent Web renovations, Random House today unveiled their publisher-driven alternative to Google: a new, full-text search engine of over 5,000 new and backlist books including browsable samples of select titles. The most interesting thing here is that book samples can be syndicated on other websites through a page-flipping browser widget (Flash 9 required) that you embed with a bit of cut-and-paste code (like a YouTube clip). It’s a nice little tool, though it comes in two sizes only — one that’s too small to read, and one that embedded would take up most of a web page (plus it keeps crashing my browser). Compare below with HarperCollins’ simpler embeddable book link:

Worth noting here is that both the search engine and the sampling widget were produced by Random House in-house. Too many digital forays by major publishers are accomplished by hiring an external Web shop, meaning of course that little ends up being learned within the institution. It’s an old mantra of Bob’s that publishers’ digital budgets would be better spent by throwing 20 grand at a bright young editor or assistant editor a few years out of college and charging them with the task of doing something interesting than by pouring huge sums into elaborate revampings from the outside. Random House’s recent home improvements were almost certainly more expensive, and more focused on infrastructure and marketing than on genuinely reinventing books, but they indicate a do it yourself approach that could, maybe, lead in new directions.