Betting Against Google’s Answer to Wikipedia

As many now know, Google announced Fri­day that it’s test­ing a new con­tent ini­tia­tive — dubbed “knol” — that it hopes will rival Wikipedia. Real­iz­ing that Wikipedia entries rank first on 27% of all Google search result pages, the folks at Google­plex could­n’t resist launch­ing a com­pet­i­tive prod­uct. In announc­ing “knol,” the com­pa­ny high­light­ed two prob­lems that this new con­tent prod­uct will address:

1) “There are mil­lions of peo­ple who pos­sess use­ful knowl­edge that they would love to share,” but they don’t share that knowl­edge “because it is not easy enough to do that.”

2) “The key idea behind the knol project is to high­light authors. Books have authors’ names right on the cov­er, news arti­cles have bylines, sci­en­tif­ic arti­cles always have authors — but some­how the web evolved with­out a strong stan­dard to keep authors names high­light­ed. We believe that know­ing who wrote what will sig­nif­i­cant­ly help users make bet­ter use of web con­tent.”

How “knol” attempts to solve these prob­lems is fair­ly straight­for­ward. It will pro­vide experts with user-friend­ly tem­plates for writ­ing and pub­lish­ing ency­clo­pe­dia entries (or “knols”) on the web. And since a pic­ture is appar­ent­ly worth a thou­sand words, I rec­om­mend that you take a look at a sam­ple screen­shot here. Depart­ing from Wikipedia, Google’s project will cater to the indi­vid­ual author, not com­mu­ni­ties of authors. And it will encour­age many ency­clo­pe­dia entries on the same top­ic, as opposed to one uni­fied text. Google then assumes that the cream will rise to the top. If 20 peo­ple craft “knols” on “string the­o­ry,” then the best one — pre­sum­ably the one that gets the most links from qual­i­ty sites — will rise high­est in the search rank­ings.

Google’s con­cept is not alto­geth­er bad. But it’s also one of the more ordi­nary ideas to come out of Moun­tain View, and I’m guess­ing that the results will fall short of cor­po­rate expec­ta­tions. Here’s why:

Most fun­da­men­tal­ly, the infor­ma­tion gen­er­at­ed by these “knols” will be sub­stan­dard com­pared to what you’ll find on Wikipedia. Although the screen­shot pro­vid­ed by Google nice­ly fea­tured a Stan­ford Uni­ver­si­ty schol­ar writ­ing on “Insom­nia,” the real­i­ty is that few experts of this stature will take the time to con­tribute. Take my word for it. I’ve spent the past five years try­ing to get schol­ars from elite uni­ver­si­ties, includ­ing Stan­ford, to bring their ideas to the out­side world, and it’s often not their first pri­or­i­ty. They just have too many oth­er things com­pet­ing for their time. More often than not, Google’s knols will be writ­ten by authors with less­er, if not dubi­ous, cre­den­tials. The mediocre entries will be many; the great ones, few. And this will leave Google’s con­tent in a weak­er posi­tion rel­a­tive to Wikipedia.

To be clear, Wikipedi­a’s over­all tal­ent pool may not be much bet­ter. But Wikipedi­a’s mod­el has an impor­tant built-in advan­tage. A com­mu­ni­ty of writ­ers focus­ing on the same text will cor­rect one anoth­er and improve the over­all prod­uct over time. The final text becomes greater than the sum of its authors. Mean­while, Google’s mod­el, which will pro­duce a pro­lif­er­a­tion of lack­lus­ter entries on the same sub­ject, does­n’t include any kind of strong self-cor­rect­ing mech­a­nism that will improve the entries. The com­pa­ny seems to think that user feed­back, name recog­ni­tion, and a share of ad rev­enue (which prob­a­bly won’t amount to much) will do the trick. But that seems like wish­ful think­ing, and I’m bas­ing that on sev­er­al years of work­ing at, which inte­grat­ed many of the same ele­ments into its mod­el. Strike one against Google.

If you’re look­ing for Strikes 2 and 3, let me out­line them briefly.

Strike 2 comes down to false premis­es: When you step back and exam­ine Google’s rea­sons for cre­at­ing project “knol,” they don’t hold up to scruti­ny. These days, pub­lish­ing on the web is fair­ly dum­my proof. Free blog­ging soft­ware, Google Page Cre­ator, Yahoo’s Geoc­i­ties and Wikipedia — these tools have made it incred­i­bly easy to pub­lish to the web. (Some­how, writ­ers have fig­ured out how to post 2,125,453 arti­cles to Wikipedia.) The argu­ment that tech­nol­o­gy is hold­ing back would-be ency­clo­pe­dia writ­ers just does­n’t fly. Nor does the notion that we’d get bet­ter qual­i­ty ency­clo­pe­dia entries if only authors could attach their names to what they write. On the one hand, anonymi­ty has­n’t slowed down Wikipedia at all. On the oth­er, many legit­i­mate experts will see writ­ing “knols” as being a slight step above “van­i­ty” pub­lish­ing, but not much more. In short, not a good use of their time.

Strike 3 turns on momen­tum and the lack of game-chang­ing func­tion­al­i­ty: Not long after YouTube launched and proved the via­bil­i­ty of video shar­ing, Google cre­at­ed its own com­pet­i­tive unit, Google Video. By the next year, Google real­ized it would nev­er catch up and bought YouTube for $1.65 bil­lion. Wikipedia, in com­par­i­son, has had a much longer head start. For six years, it has been refin­ing its mod­el, grow­ing traf­fic, and gain­ing user loy­al­ty. That’s a sub­stan­tial and most like­ly insur­mount­able lead. True, once upon a time a young Google came out of nowhere and knocked an estab­lished Yahoo out of its lead­er­ship role. But that hap­pened when Google brought its game-chang­ing search tech­nol­o­gy to mar­ket. With “knol,” how­ev­er, there’s no such game-chang­ing tech­nol­o­gy on dis­play — noth­ing that sub­stan­tial­ly changes how knowl­edge gets cre­at­ed. Google and its engi­neers cer­tain­ly excel at man­ag­ing knowl­edge and pro­duce many great prod­ucts (for which I’m per­son­al­ly thank­ful). But get­ting into the knowl­edge cre­ation busi­ness may pose new chal­lenges, ones that will require the Google staff to go beyond algo­rithms and think­ing in terms of 0s and 1s.

Sub­scribe to Our Feed

by | Permalink | Comments (13) |

Sup­port Open Cul­ture

We’re hop­ing to rely on our loy­al read­ers rather than errat­ic ads. To sup­port Open Cul­ture’s edu­ca­tion­al mis­sion, please con­sid­er mak­ing a dona­tion. We accept Pay­Pal, Ven­mo (@openculture), Patre­on and Cryp­to! Please find all options here. We thank you!

Comments (13)
You can skip to the end and leave a response. Pinging is currently not allowed.
  • One notice­able thing about the mock­up graph­ic is the promi­nent Cre­ative Com­mons CC-by 3.0 logo. This caught our eye at Wikipedia. The point of Wikipedia is not in fact to run a hideous­ly pop­u­lar (and expen­sive) web­site, but to cre­ate a body of freely-reusable edu­ca­tion­al con­tent. So IF, I say IF, Google require Knols to be under a prop­er free-con­tent licence, then that’ll be a big win for every­one.

    (And if they don’t, they’ll just be anoth­er or Yahoo Answers. Or Google Answers — remem­ber Google Answers? I bet Google does.)

  • John says:

    Ha. Get­ting a lit­tle dig in are we? Just kid­ding. Nice to see the blog evolv­ing.

  • Barry Kort says:

    It’s hard to beat Wikipedia as a com­pendi­um of pop­u­lar cul­ture, since it’s writ­ten by a pop­ulist com­mu­ni­ty that rev­els in pop­u­lar cul­ture.

    But when it comes to schol­ar­ly sub­jects, Wikipedia suf­fers from a juve­nile cul­ture that is hos­tile to the aca­d­e­m­ic mod­el.

    Author­i­ta­tive sub­ject mat­ter experts are unwel­come on Wikipedia, where the dom­i­nant play­ers eschew con­sci­en­tious schol­ar­ship in favor of dra­ma, con­tro­ver­sy, and gam­ing the sys­tem for fun and prof­it.

  • Dan Colman says:


    I per­son­al­ly haven’t had the same expe­ri­ence with Wikipedia. True, you can find entries on fair­ly inane top­ics. But it cov­ers pret­ty much any area of true schol­ar­ly inter­est, and often the con­tent is quite sol­id.

    Here I’m remind­ed of a New York Times arti­cle that report­ed on how Wikipedia com­pares to Brit­tan­i­ca when deal­ing with sub­stan­tive sci­en­tif­ic top­ics. I quote:

    “A study last month in Nature showed that the deci­sion is far from clear-cut. Call­ing on experts to com­pare 42 com­pet­ing entries, the jour­nal count­ed an aver­age of four errors per arti­cle in Wikipedia — and three in Bri­tan­ni­ca. That is not much of a dif­fer­ence, and a look at the details only adds to the anx­i­ety.”

    A cou­ple oth­er rel­e­vant quotes to add:

    “What­ev­er their short­com­ings, nei­ther ency­clo­pe­dia appears to be as error-prone as one might have inferred from Nature, and if Bri­tan­ni­ca has an edge in accu­ra­cy, Wikipedia seems bound to catch up.”

    “It seems nat­ur­al that over time, thou­sands, then mil­lions of inex­pert Wikipedi­ans — even with an occa­sion­al sabo­teur in their midst — can pro­duce a bet­ter prod­uct than a far small­er num­ber of iso­lat­ed experts ever could.”


  • Carol A says:

    I sup­pose any com­mu­ni­ca­tion method which caters to every­one is going to attract some triv­ial con­tent — think of TV and radio! But at least on the inter­net you can ignore the rub­bish. Wikipedia does seem to pro­vide the bet­ter method, but to give Google their due, their book search engine was a joke when it first start­ed, now I find it has evolved into some­thing quite usable. Per­haps they do lis­ten to crit­i­cism?

  • Dan Colman says:

    Hi Car­ol,

    I agree. I would­n’t be at all sur­prised to see Knol evolve and turn into some­thing quite dif­fer­ent over time. Google is clear­ly a bright and respon­sive com­pa­ny. I was more just respond­ing today to what we see right now.


  • Jon Awbrey says:

    The notion that “a com­mu­ni­ty of writ­ers focus­ing on the same text will cor­rect one anoth­er and improve the over­all prod­uct over time” or that “the final text becomes greater than the sum of its authors” is sad­ly out of keep­ing with the real­i­ty of Wikipedia, where arti­cles cre­at­ed by knowl­edge­able authors are more like­ly to be degrad­ed over time by hordes of inept users and pow­er-trip­ping admin­is­tra­tors who nei­ther know nor care any­thing about the sub­ject mat­ters in ques­tion.

  • I think Wikipedia and Google Knol can co-exist, and I do not entire­ly agree with your argu­ment that there will be too few experts will­ing to write a knol.. there are some incen­tives (both mon­e­tary and rep­u­ta­tion­al) to do it, even for a high-pro­file expert. Espe­cial­ly because it is Google, and not some unknown start­up, a pos­si­bly larg­er audi­ence can be reached.

  • Sniperz11 says:


    I’ve been a reg­u­lar Wikipedia edi­tor for a year now, and was an anony­mous edi­tor for a year before that.. I havent had a chance to see Knol in action, but based on what Google is say­ing, I have a few obser­va­tions:

    1. Col­lab­o­ra­tive page build­ing is FAR FAR bet­ter than get­ting ones name rec­og­nized. I’ve built pages togeth­er with oth­er edi­tors, some­times, work­ing live by chat­ting on IRC and build­ing togeth­er. Believe me, the con­tent that comes out of syn­er­gy is incom­pa­ra­bly bet­ter than when done alone.

    Theres anoth­er issue with work­ing alone. After some time, bore­dom sets in, as it usu­al­ly does when work­ing in a vac­u­um.

    2. Self-sat­is­fac­tion trumps attri­bu­tion: This is anoth­er major hole I find in Google’s log­ic. If you look at Wikipedia, almost half of all the major edits are made by anony­mous users, and almost 95% of the 6 mil­lion reg­is­tered users, includ­ing me, use a pseu­do­nym to con­tribute.

    Thats an indi­ca­tion that the con­tributers are doing it for some­thing more than recog­ni­tion, which, frankly, no one cares about.

    4. As Jim­bo Wales observed, with the Knol, the oppor­tu­ni­ties for spam­mers is end­less… imag­ine the num­ber of entries for Via­gra one will get.

    5. Rank­ing: This is anoth­er con­cern — most read­ers will be drawn to the con­tro­ver­sial or blog-like entries, which means that actu­al good entries will get pushed down in pri­or­i­ty.

    6. Com­mu­ni­ty: This is the BIGGEST advan­tage that Wikipedia has… if you look at any of the over 600 Wikipro­jects, which bring togeth­er edi­tors to improve pages in their area of inter­est, there is a feel­ing of com­mu­ni­ty bon­homie that exists that makes the out­put work bet­ter than when done in an ad-hoc man­ner. And with the lack of such an orga­ni­za­tion­al set­up on Knol, that will be absent. So while con­tent will be huge, qual­i­ty will most like­ly be abysmal.

    Over­all, I think Google has start­ed Knol just because it sees a com­mer­cial oppor­tu­ni­ty there, and as anoth­er means of it want­i­ng to con­trol the Inter­net.

    The pity of the whole sit­u­a­tion is that Wikipedia entries will be engi­neered down in the Google rank­ings by changes to their Algo­rithm. Still, look­ing at how the Wikipedia Search page gets over 16 mil­lion vis­i­tors dai­ly, that may not be a big dis­ad­van­tage.


  • […] in Decem­ber 2007, I made a bet against Google Knol, the search giant’s answer to Wikipedia. In a fair­ly involved piece, I list­ed three rea­sons […]

  • Monitors says:

    Hi there. Today the appli­ca­tion of win­dows live mes­sen­ger. That is the def­i­nite­ly tac­ti­cal­ly cre­at­ed write-up. I shall be absolute to book mark them obtain­able back to con­tin­ue read­ing of one’s use­ful facts. We appre­ci­ate you your write-up. I’ll def­i­nite­ly come back.

  • Foreign movies says:

    Woah this web­site is def­i­nite­ly superb i enjoy read­ing your con­tent reg­u­lar­ly. Remain on the very good pic­tures! You rec­og­nize, a great deal of peo­ple are seek­ing around just for this infor­ma­tion, it is pos­si­ble to guide these indi­vid­u­als con­sid­er­ably.

  • Corel says:

    Amaz­ing whip! I wish to begin­ner even though you amend your site, how could we signed up for a blog site? The actu­al pro­file assist­ed us a suit­able cope. My part­ner and i end­ed up a bit famil­iar of this the trans­mit avail­able sparkling clear thought

Leave a Reply

Open Culture was founded by Dan Colman.