Broker translation memory? Why not? Entrepreneurs long ago found that you could broker anything as long as you had willing sellers and buyers. The internet lets realtors, commodity traders, financial brokers, and now TMMP bring these buyers and sellers together in worldwide marketplaces. The middlemen benefit by collecting a commission on each sale.

But will brokering translation memories work? We see a long slog ahead for the principals, largely due to the challenge of getting companies to think more strategically about linguistic assets such as translation memory. With 90% of translation outsourced, most companies rely on their language service providers to manage their TMs. In the best of worlds, the LSPs use a single, unified TM. In the real world, companies use several LSPs to do their work and don’t often require them to share TMs. Of course more thoughtful potential licensors will claim that their TMs have strategic value as intellectual property, but anyone who wants to take the time can review a company’s English and Albanian documentation and reverse-engineer a TM.

TMMP provides a basic framework for an active marketplace. Before they find buyers lining up, TMMP has to convince enough buyers to offer up their TM crown jewels for licensing. Potential licensors will have to pinpoint their TMs and their locations, retrieve them, remove any proprietary terms like “flux capacitor” that identify unreleased products or future features, and certify quality. For now, TMMP requires a licensor’s assertion of basic quality attributes such as editing and verification.

Will TMMP succeed? Getting some giants like Airbus, General Motors, Microsoft, SAP, and Sony to offer up their TMs could jumpstart this effort, but it’s more likely that these companies will first offer them to their own supply chains. It’s up to TMMP to demonstrate why it should be the “honest broker” in TM transactions.