Crowdsourcing a Catalog

Last Wednesday, I asked for help selecting images that I should consider launching my upcoming print store with. The response has been awesome. Out of the collection of almost 6000 images in my photo stream, 129 were tagged with print requests by a virtual army of people. This has been crowdsourcing at its best.

Even better, this exercise has resulted both validation and surprises. There were several photographs that I knew would be good enough to make the cut, and having them also picked by my readers is sweet validation. But, there were many more photographs picked out that I might not have selected on a first pass cut myself. And many of these photographs picked up a number of printmeplease1..n tags indicating that the support was coming from more than one person.

Throughout the process, I was keeping tabs on my Flickr account watching the ebb and flow of tags go by. The process of watching this happen was instructional. As well, the printmeplease1..n tags have turned out to be a great tool to use to measure response with. It's not a perfect, by any means, but the information provided by people who left them is incredibly useful.

Sedona Reflections

The hands down winner was the photograph to the right. It was made in Sedona, Arizona with the help of a puddle midway up Bell Rock. It scored all the way to a printmeplease13 tag. The next closest images, two of them, scored printmeplease8 tags. With that kind of feedback, you can be absolutely sure that this photo will make the cut and be in the store.

I did receive some comments from readers indicating that the sheer scope of my request was daunting. After all, 6000 images are quite a few to look through. I was actually a bit worried that this would put some people off, but despite that hurdle and incredible number of people dug through and helped out. I owe a debt of gratitude to all of you. You've all truly helped out in this process. Everyone who left a tag should get a FlickrMail before the store launches with a special thank you.

Now, I've got my work cut out for me. Over the next few days, I'll be sitting down with incredibly valuable feedback, making test prints, and selecting the photographs will make the cut for the first dozen or so that will be offered as prints in the store. I don't think I'll stop there, however. With the amount of feedback I've received, there will probably be several waves of a dozen each that will be added to the store over the first few months of its life.

If you're interested in what people selected, here are all the images that were tagged:

White Sands Boiler Bay The Waffle Jenny Lake Boogie Amsterdam Door to Nowhere Amsterdam Bison Bowman Lake Casa Batlló Stairs Tim O'Reilly Marine Layer Rainbow View from Many Glacier Lodge Kloster Eberbach San Sebastián Street Very Large Array Sundial Bridge Pylon Golden Gate Polebridge Montana Sunset Very Large Array Tree and Sky Lake McDonald Cherubs Boogie at the Window Dead Tree Tenaya Lake Crucifix Norris Geyser Basin Mt Shasta on a Cloudy Day Sedona Reflections Stations of the Cross Sand Dunes Leaves Sand Dunes Very Large Array Cupcakes and Sculpture St Mary Lake Golden Gate Pismo Beach Chrysler Building View Sunrise over the Pearl Sunset Mammoth Hot Springs Oxbow Bend Amsterdam Wind Sculpture Windblown Mt. Shasta Sunset Reflections Grand Canyon and Clouds Norris Geyser Basin Haystack Rock Bumpass Hell Big Round Cubatron White Sands View from Many Glacier Lodge Street San Andreas Mountains The Man Starts to Burn Crater Lake San Francisco Bay Night View Graffiti Serpent Mother Mammoth Hot Springs Mono Lake Basin Bryce Trees Casa Batlló Walls Sunset at Westgard Pass Feet Road out of Death Valley Sedona Rocks Mono Basin Poconip Northern California Wetlands Gower Peninsula James Duncan Davidson Park Avenue Boats RailsConf Europe Audience Gower Peninsula Rabbit Rocks Waiting for the U-Bahn Sunset Southwest Colorado 50th Street and Park Ave Beach at Gower Peninsula Hope Mother Eastern Sierra Dust Wind and Sand Bryce Canyon Hoodoos Placa Catalunya Fountain San Sebastián Street Black Rock City Heceta Head Great Fountain Geyser Reflections Norris Geyser Basin White Sands I.T. Reflected Sunset Sunset at Boiler Bay Boiler Bay Wave Mules on the Trail Crater Lake Bikes and Lasers Hotel Atrium Heceta Head Oak Creek Sunset Friedrichstraße Station Das Kurhaus Fountain Grand Tetons Museo de Ciences Principe Felipe Poconip Rime Peppers Steve Ballmer Forest Meets Field Pedestrian Bridge Sushi Bar San Sebastián Street Oxbow Bend Park Place Cannon Beach Cannon Beach Northern California Wetlands Thermal Pool Snowy Trees Delicate Arch White Sands

Again, thank you to all who participated and helped me out. Not only have you helped to shape the catalog, you've given me the knowledge that I can proceed on with this endeavor.

Related Posts:

0 TrackBacks

Listed below are links to blogs that reference this entry: Crowdsourcing a Catalog.

TrackBack URL for this entry: http://duncandavidson.com/cgi-bin/mt/mt-tb.cgi/20

4 Comments

Nice one Duncan!

Who's the saucy Flickr user that would like your self portrait on their wall, then? ;)

That's a question I'll leave unanswered. But it's highly unlikely that I'll be putting that one up for sale in the store. I dunno, but I think I'd feel weird about that.

The sheer number was pretty daunting. I started somewhere near the beginning, tagged a couple then got sidetracked. I intended to come back, but didn't. I'm curious if the photos at the beginning of your sets were more heavily tagged.

I think if you (or anyone) were to do this again, it would be helpful to have a little app that displayed random images from a set (or sets) of photos, and let people comment/tag from there.

Brad: There is some weight—probably quite a bit really—to the tagged photos being towards the more recent part of the timeline than not. As to whether user fatigue contributed to that, it's hard to know. I'd be pretty sure that some amount of that weighting would be because the more recent photos tend toward being better as well.

I agree that some sort of randomization function would be good, as would the ability to filter out entire categories of photos. For example, a big bulk of my photostream is composed of conference photos would would be nice to filter out.

Thanks for your help!

Leave a comment

Recent Entries

Why Auto White Balance Isn't Perfect
Every digital camera made comes with a promise of making perfect pictures with the click of a button. The algorithms used to deliver on this promise, however, are by no means perfect, including the auto white balance algorithm.
Print Store Update
On April 2nd, I was pretty optimistic that I was just a couple of weeks away from launching an online store with the first of my prints for sale. As you know, I asked for a bit of help from all of you to help out. That process turned out to be a fantastic experience and resulted in a lot of feedback.
Ignite SF April 2008
As part of Web 2.0 Expo SF 2008, Brady Forrest and company hosted an Ignite SF at the DNA Lounge last night. Ignite nights are always a good time. What's Ignite? Take 14 or so speakers, give them 5 minutes each to talk, and see what happens. The result is almost universally fun. And, thanks to the format, if somebody's talk is a dud, well, you only have to wait 5 minutes for the next one.