Your Two Cents

The Denver Post’s article commenting functionality has undergone many changes since it debuted in 2007. We have had more than one million comments posted. Our moderators have deleted more than 25,000 comments. Hundreds of liberals and conservatives have protested our deletions, claiming that we only delete conservative/liberal comments.

While the traffic our comments section generates has been sort of okay — about 3.5% of overall site traffic — ads are not enough to support the costs. You’ve heard about newspapers’ difficulty figuring out an online ad revenue model. Well, this week, we figured it out. We call this the “Your Two Cents” plan.

Your Two Cents
Part of the challenge with our article comments is that commenters write too much. Another challenge is many people don’t use their real name, and those people tend to be the ones who also go on profanity-laden hate-parades. The other part fo the challenge is we, The Denver Post, aren’t making enough money with our article comments. Your Two Cents addresses all of these challenges.

Starting next week, April 5, The Denver Post will start charging commenters for each word in each comment they write. Two cents a word. This means:

1. Commenters will have a financial incentive to be concise
2. Commenters must provide real first and last names in order to pay for their commenting
3. The Denver Post will increase online profits, allowing us to invest in comment moderation and other online endeavors

If you have any thoughts, suggestions or praise for this plan, please share it in the comments below.

1. Provide context about the exact place that I live. Context makes information actionable.
2. Make accessible and linkable historical information about the place that I live. News sites are a community resource — time to start acting like one.
3. Give news sites exponentially more entry points to the information they’re already publishing. More entry points makes information more findable.
4. Make local political news and information more accessible. This makes politics more approachable and actionable to those not already disposed to follow it.

1. An index of all the facts included in the articles they publish

This means a list of facts, as well as a means to link directly to the part in the article that fact exists.

Example: McDonald’s buys more than 3 billion pounds of potatoes annually across the globe. This nugget of information is more interesting than the article’s headline, McDonald’s seeks better ‘tater for its French fries, yet it was left embedded in the article body for only the most curious to find.

Indexing facts does more than provide new and engaging entry points to existing content. Facilitating easy citations with facts and links to facts can improve the quality of conversations on news-site article comments, and it can also encourage wikipedia users to cite the news site with the tools that make it easy to cite.

2. News archives. Not just from the last month — from the last year, ten years, fifty years, century.

Every local news-dot-com publishing with a newspaper is sitting on a goldmine of archived content. The New York Times hired reCaptcha to help digitize their archives — sure, the NYT’s web strategy doesn’t always align with that of local news-dot-coms, but in this case, they’re onto something.

3. Indexes of news and information by zip code

Denver’s a decently big city. We’ve got 72 neighborhoods and xx zip codes. If there were a place I could go to get all the news, calendar events, and classified listings in my zip code, I would. Not only that, I would tell my neighbors about it. Indexing by zip codes gives a hook for loyal readers to introduce your site to the people that live around them that may not care for your publication, and it gives the non-loyal readers, the non-news junkies a compelling reason to visit.

4. Indexes of information on local politicians, organized by politician.

I don’t care about your catch-all “local politics” category. I care about about the politicians that represent me, and I want an easy way to find out everything they’re doing. That means not just local politicians either — that means the people repping me in the statehouse, my U.S. House representative and my U.S. senator.

Looking at “local” as a catch-all bucket rather than a collection of specific and distinct pieces is a superficial approach to publishing.

The New York Times’ Represent application approaches local politics in a mature and fully fleshed manner.

Here are some examples of catch-all local politics buckets:

• Denver Post: Local Politics
• Boston Herald: Local Politics
• Dallas Morning News: Local Politics

5. Indexes of major crimes, by date, with crime stats aggregated by month, year and every type of location that’s available (county, zip code, neighborhood, street, block etc.).

Yes, this is the type of information you see Everyblock and Adrian Holovaty pushing online. I’m not saying publish data-driven presentations of all crimes — I’m saying start with the big ones, see how that works, and go from there. Publishing per-capita rates for violent crimes opens a window on urban vs. suburban living, on what’s happening in the places we call home and work, and how these incidents trend over time.

I’m going to repeat that: How these incidents trend over time. Crime drives a large part of the news truck, but so often it’s crime without context. Now that local news is online, it has the opportunity to give context to the information it publishes. What would this context do? Turn crime news from the hand-wringing / rubberneck activity and make the crime information actionable. If arson has increased 200% in my zipcode (80204) in the last year, that’s worth asking my police department and local government about.

More meaningful metrics would translate visitor interest, disinterest and loyalty into numbers that can be viewed as a whole and within the context of particular site content types, classifications or products (home page, article page, sections, photos, photo galleries, data ghettoes, etc.).

Even more meaningful metrics would measure all of the above among visitors from the particular local news site’s circulation area. In the case of my employer, that’s specifically Denver, and generally Colorado. Forty percent of our site visitors come from Colorado, and that number’s rising, which is good. But those numbers don’t tell what the churn rate is, or what percentage of our Colorado visitors are repeat-daily visitors, or how the repeat-daily number has changed over time.

This is an incomplete list of meaningful metrics for a local online news site, written in the context of a Denver Colorado online news site:

• What non-branded search terms are Coloradans using to find our site content? Once they come, who stays for a second click, and what search terms result in the most second-clicks? What search terms result in the fewest second-clicks?
• What’s the return frequency of Coloradans? Among Denver residents? Among Aurora residents? How has that number changed over time?
• Which sections (news / sports / business / entertainment) have the highest percentage of visits from Coloradans? Are any sections declining in that number? Is that decline a seasonal issue or is it longer lasting?
• What’s the bounce rate among visitors who enter at articles? Does that change based on the section the article’s in? Does that change based on whether it’s a Colorado visitor? How has that rate changed over time?
• What’s your homepage bounce rate among visitors who arrived at your homepage for the first time today? What’s the rate among visitors who have already visited your homepage today? How does that number change over time?
• This is something that is more difficult to measure: How many readers make it to the end of an article? Some javascript that hooks into the y-position of the last paragraph and measures that against the scroll of the window would be necessary to get into this metric, and even then it wouldn’t be wholly accurate (the bigger the browser window, the shorter the article, the less accuracy)

Note: All of these metrics, and any metrics at all, become much more useful when keeping a site log of incidents (breaking news, special projects) and site changes.

Got any more metrics to add? Share ‘em below.

Two weeks ago I was talking with Dan Petty (our online intern at the Denver Post) about a Drudge Report link to our site from the week before. That link, and the words used in that link, had made this denverpost.com article a top-3 result on Google if you were to search for people shouting.

Another article on our site, Debunking Canadian health care myths, has been getting steady traffic, fed by links posted every so often on trafficked site, since June 7. I was curious how it placed in searches for Canadian health care. It wasn’t a first-page result. However, a CommonDreams.org page that had cut and paste (i.e. stolen) the Denver Post article in its entirety was the #7 result. Good for them? Not so much, though the Post is not the only ones who get ripped off by Common Dreams.

Anyway. I contacted Common Dreams about this, and they were nice enough to trim their copy of our article down to five paragraphs. And what happened next is what I find the most interesting bit of this: Within two days of Common Dreams trimming down their article, the Denver Post’s version of Debunking Canadian health care myths was the #7 result in Google in searches for Canadian health care. It’s not anymore, but it was.

What this means:

• Google’s algorithm tries to figure out the original source of an article, and just because a cut-and-paste article links to the original doesn’t put it out of the competition. If you have the full article, and you have more links to your version of the article and a higher page rank, Google will likely think that you’re the source. Despite a Common Dreams link to the Denver Post’s article with this text “Published on Sunday, June 7, 2009 by The Denver Post,” Google still decided to post the Common Dreams article in the first-page results.
• Wholesale cut-and-pasting of your content is probably worth addressing. Well, sort of. If you’re a local newspaper, those eyeballs from outside your circulation area aren’t so valuable to your advertisers. They will also skew your numbers (see Gerry McGovern’s Volume is the wrong way to measure web success. However, taking search engine ownership of phrases central to your coverage is a goal often overlooked by local news orgs, and addressing copy-and-paste-cats is a prong of a healthy search engine strategy.

If you have any real-life examples of Google’s search algorithm at work, do share.

Reading this followed my discovery of the Kankakee Daily Journal’s site, which allows comments on the two or three grafs of articles they provide online — the rest of the word-based news is tucked behind the pay e-edition wall.

I’m asking myself: Is PDF-style delivery really the future of newspapers? The e-edition PDFs are attractive to newspapers — they look like the print edition, and they contribute to print-circ numbers. But they also ignore the possibilities of online advertising, which is a glaring problem with many newspaper-dot-coms.

1. Build an audience around a link-based social news site for local information.
2. Once your site has some power users, give them blogs.
3. Team up with hacker/developers for special projects.
4. Once your site has built some momentum, hire a part-time ad sales person .
5. Once your site’s earning $300-$400 a week in profit, start contracting with freelance journalists. Scoop local, stuck-in-the-print paper.
6. Add more writers to your blogs.
7. Build enough audience so a front-page link on your site will deliver at least 1,500 clicks to its destination, your blogs are breaking news that isn’t anywhere else, and you have the ability to set the agenda in the community you cover.

1. Hyperlinking the Real World – ReadWriteWeb: “Welcome the dawn of the age of ubiquitous computing! Just wait until everything is rfid tagged so you can google search for your socks.”
2. $200 Laptops Break a Business Model – NYTimes.com
3. Five fatal flaws in local Internet ad sales
4. Web Usability Blunders That Still Piss Me Off
5. You Will Probably Fail in a Boring and Project-specific Way

Get these links delivered via twitter here, or subscribe to this online journalism link feed (low-frequency, 3-5 posts a week) via rss / web feeds here.

So it brought me some pleasure to stumble on this photo, taken of the desk of one of the Rocky’s online guys, who had a message about the Rocky’s soon-to-be former corporate parent, the E. W. Scripps Company:

From one of their (short-time) onliners' desk

When I visit PressDemocrat.com, I go for one thing: Sonoma County news. Someone in Mendocino County might visit the site for Mendo County news, which is great, but not the reason I visit. Ok, with that said, how do I locate Sonoma County news on PressDemocrat.com. Ahh… herein lies the problem. Local news granularity is sorely missing on the site.

When scrolling down PressDemocrat.com’s frontpage, you won’t find sections for “Santa Rosa news” or “Windsor news”

When I worked at the Winston-Salem newspaper, we had sections for each of the counties we covered, and I’m pretty sure we had feeds for each of them. One of our managers suggested getting rid of them — the traffic wasn’t particularly high to any one of the sections…. however, if you added the traffic to all of the county-based sections together, it was traffic worth considering.

And if you like that anecdote, I’ve got another. Denver’s Rocky Mountain News, despite their “Closer To Home” slogan and (some say) general reputation for being more focused on the local, has no place-based online sections or place-based RSS feeds. The Denver Post, my employer, has both. The Post had place-based RSS feeds for the Denver Metro area before I started working there in October 2006.

I don’t get it — you’re a local news organization, you publish most of your information about a specific place on this planet, yet you do nothing to highlight, filter, or organize your place-based information? These city- / county- / neighborhood- / street- / block-based news feeds are just the tip of the location-based information iceberg. What can you do with a dateline? With a locator map? With a photo? A crime blotter? A classified ad? A banner ad? A calendar entry?

I wrote about some of this at the beginning of 2008 in this blog post, Three ways that online changes the “Where?” question, journalistically. I’m working on answers in my day-job, some of the time. If you’ve got ideas for a non-day-job project related to this, let me know.

The Onion’s making a push into the online local entertainment markets of ten cities — Madison, Milwaukee, Chicago and Austin already, and Los Angeles, New York, Washington D.C., San Francisco, Denver and the Twin Cities to come, says The Onion’s press release from PR Newswire.

They’re changing the name from AV Club to The Decider. You can see Chicago’s “Decider” site live here. Notice the decider.com domain — why they decided to use a new domain name and not cash in on the existing search-engine cred of theonion.com I have no idea.

So, beside the news that there’s another face in the local-entertainment-guide information game, the other interesting bit (which I heard from a friend of mine) is The Onion is taking all the market-specific articles they’ve published in the AV Club’s archives, and publishing them anew online. The advantage there is that these sites launch with more than just a front page and a handful of articles… also, it’s more information for search engines to index… also, and this is just me thinking out loud, I wonder why other print media with extensive archives don’t take advantage of that gold mine of existing, low-overhead, local information?