My experiences with freebase
This article is about Freebase, the crowd-sourced semantic database.
The fate of Freebase, the crowd-sourced semantic database, is anissue that has been largely overlooked when it comes to the accuracy of Google’s search returns. I’m afraid this isn’t a confession about hardcore drugs use, so look elsewhere if that’s what you were hoping for.
Given how late to the party I am on this one I’m going to assume that any joke worth making has already been played out, so please try to imagine I came up with a highly amusing and subversive gag about how freebasing is also a drugs thing. Maybe I mentioned the late Richard Pryor. We all laughed, probably rather awkwardly.
Freebase and its impact on Google Search
While carrying out a frustrating piece of work with Google My Business in 2014 I learned that the Knowledge Graph was heavily influenced by the information held on Freebase. I was working for a university at the time, and they typically have very complicated online footprints that are a challenge to keep up to date.
I was already familiar with the Knowledge Graph (introduced in 2012) but I’d never heard of Freebase — and I suspect I’m not the only one — but it turns out that this collaborative knowledge base has quietly been playing a key role in informing Google Search returns since they acquired it in 2010.
Some of the facts about my employer were outdated or just plain wrong (most noticeably their name, which was a bit of a howler). These inaccuracies were influencing the knowledge card results Google presented for users searching for the university, and causing genuine confusion. Thankfully the entry was easy to amend after signing up for an account, and the changes gradually appeared in search results alongside other improvements resulting from tidying up the institution’s My Business data.
At the time I speculated that Freebase wasn’t the target of any serious investment, and had a sense that the platform was largely overlooked in the Google Search roadmap. I was right on the first point, but definitely wrong on the second… it turns out that the plan was to kill it off.
I was recently asked to take on a similar piece of work for another university, but it turns out a lot has changed in the last couple of years. In December 2014 Google announced a plan to discontinue Freebase, although the plan has changed a few times and I think users have been left a little in the dark. Unfortunately, it’s worse than that for anyone trying to correct or update data — the site has been read-only since 31 March 2015.
Two months from now it will have been locked for a year, and with ~58 million topics and ~3.17 billion facts there’s an ever growing margin of data rot that concerned parties haven’t been able to remedy. It’s also something that will only get worse the longer it goes on. The website was due to be retired in June 2015 but it’s still up and running, read-only warts and all.
Anyone trying to manage how their company or institution is referenced on the web should definitely keep an eye on what Google are doing here (and the pace at which they’re doing it).
Freebase was a messy folksonomy and a noble concept, but its replacement is currently a messy movable feast. I know that Google wants to get us all to a better place with its Knowledge Graph but in the meantime users with inaccurate Freebase data are in limbo.
I wanted to finish with a drugs-related joke as a callback to the beginning of this article, something Richard Pryor might have appreciated. I just couldn’t think of anything.
- Official Google Blog | Introducing the Knowledge Graph: things, not strings (16 March 2012)
- Freebase announcement on Google+ (26 March 2015)
- Freebase on Google+ update (16 December 2015)
- Neil Patel: The beginner’s guide to the Google Knowledge Graph (30 June 2015)
I specialise in navigating the challenges faced by complex institutions in this digital age, driven by a restless desire to see them wield technology to achieve their true potential.
Formerly the Chief Digital Officer at one of the world’s leading business schools, I have over a decade of experience in high profile roles at the University of Cambridge and King’s College London. With a background spanning digital strategy, corporate communications, and the management of technology projects, I now consult widely across a range of sectors.
This post used to be called ‘Six things I learned about freebase’ but I couldn’t bring myself to write clickbait, and I’m not sorry.