One good piece of feedback I’ve heard about on-line communities recently is that in today’s world of easy on-line publishing. A world where everyone is now getting comfortable with on-line interaction is a world where the number of people who “shoot first” with their questions before searching has grown exponentially. We can’t change those people.
“Lurk before you leap died years ago” said several grizzled public newsgroup veterans when I asked for their feedback on technical “how to” communities. This leads to an unbelievable amount of question duplication that puts the burden on the limited number of answerers.
Though you may not have realized it I believe this is what has led to several question experts jumping off the NNTP ship and moving to on-line forums where they can moderate and start to control the flood of duplicate questions. (In the last 2-3 years the number of unique answerer types has been cut by over 20 % in our public developer newsgroups). But forums aren’t enough. Moderators and experts still can’t keep up with question volume on most sites well enough to reduce duplication. I have several ideas on how to make moderation more efficient, but those don’t get to the root cause… the duplicate questions.
I’ve been doing a lot of thinking about what more we can do to head off duplicate questions. I think I’ve got an idea that strikes a good balance between forced search first and freedom to post. Let me know what you think.
At first I was thinking of some work flow that has the user entering their new question, then doing a search, then asking them if any of the results was an answer. If a result was an answer then we ask them to cancel their post.
The problem is that this feels like a loss even in the case where users find an answer because they may have spent 30 minutes entering their whole question. There are also lots of complications with the work flow because you have to keep track of their question even if they open a search result thread.
So what we have to do is bring back search results dynamically on the new post UI page. Here is the work flow I’m imagining and some pictures to illustrate it.
1. User clicks new thread on a forum.
2. On the new post UI we bring back a heavily cached set of FAQ (sorted by most viewed in the past X days) for that forum in a search results pane on the side of the new post UI. (See figure 1)
a. This should head off the biggest dupes. Imagine heading off all of the “Activation” issues in the express forums for example.
3. The user enters a post subject
4. When the user takes the cursor out of the subject we use scripting to call a search on the subject they’ve entered in the background while they may have started entering a message body.
5. Search results are returned in the search pane based on the subject to replace the FAQ Views.
6. Users can open up the search results in a new window so they can check them out without having to leave the new post UI.
7. Either users find the answer or they make the new post. Clicking on one of the threads in the search results pane will open in a new window so the user doesn’t lose their question text.
This should be very possible with the trendy “Ajax” scripting technique used by Outlook web access web mail, google maps, etc. The goal really is not to interrupt the user who is posting, but gently remind him there are good search results.
Here are a couple of ROUGH mock-ups to help explain the idea. (Warning… really bad PM art ahead.)
Figure #1 – Raw new post UI with top forum FAQ
Figure #2 – New post UI after search completes on the subject entered
Note: In real life the three lines (posted by, post date, and forum name) would likely be better off cut to enable room for more results or more abstracts.
Comments, questions, suggestions?