Blog Post

The Problems With EveryBlock

Like it or not, hyperlocal is one of the buzzphrases in online news these days, and the site of the moment is EveryBlock.com. Founded two years ago, it was sold last week to MSNBC Interactive. Unlike other sites that provide their own local news, EveryBlock, which operates in 15 U.S. cities, automatically aggregates information from a wide list of municipal sources (e.g. 911 dispatches and land-use records) as well as news sites.

It strives to better answer the question: “What’s happening in my neighborhood?” But does it? In the wake of the sale, I decided to give the site a run-through to see whether I could rely on it as a primary source of news about my neighborhood, Capitol Hill in Seattle. Visitors to EveryBlock are asked to select their city and then to type in their neighborhood or ZIP code. They are then greeted with a long list of updates ostensibly from their neighborhood, grouped in categories, such as locations mentioned in the media, 911 dispatches, restaurant inspections, and even photos uploaded on Flickr. There’s also a spiffy map to visualize where the entries have taken place.

But during my week of testing, I found that EveryBlock had some real gaps. Often the biggest problem was that important information was either not on the site or else was presented incompletely.

For example, a top item on EveryBlock earlier this week was a piece of news pulled from the Seattle Times with the headline, “Civic Leader Frederic Danz Dies at 91.” But the only apparent connection to my neighborhood was that the funeral was taking place at a nearby funeral home.

Meanwhile, news that a nearby bar’s opening had been delayed never even made it onto the site — even though another neighborhood blog had a story about it.

Same situation with the local primary election here last week even after results were broken down by district.

There’s a reason why readers are left with these frustratingly nonexistent or incomplete pieces of information. Presumably, the only reason that Mr. Danz’s death made my page is because EveryBlock scans addresses in news articles to choose the updates it surfaces rather than paying much attention to the events themselves, which, of course, is difficult for any automated system to do.

Similarly, with the nearby bar’s opening, it may not have made the site because the original blog post only mentioned the name of the street and didn’t include an intersection or specific address.

As for the omission of the local election results they may have been too broad for EveryBlock, since a district encompasses several neighborhoods.

Updates from public data sources, meanwhile, are often devoid of context. EveryBlock prominently lists nearby 911 calls—potentially helpful information. But because the entries are dependent on the automated information that the city provides, there are no details on what actually ended up happening at a certain location — and no way to sift through which ones are actually important. An example: “Alarm bell. Engine 25 dispatched at 8:56 am on Aug. 23 at 1728 E. Madison.” Now what?

Other public updates, like recent “land use” changes, mention that a “decision” has, in fact, been made about a building plan but don’t say what that decision was.

In a blog post when it announced the sale to MSNBC Interactive, EveryBlock conceded that it is still a work in progress. It said its “current incarnation is only about 5 percent of what we want to do with it.” And it’s obviously not easy to provide complete neighborhood news without reporters on the ground or even editors sifting through the updates that are automatically generated.

But the bottom line is that in its current state, EveryBlock works better as an add-on — a place I might want to turn to if I’ve already seen the local headlines for my neighborhood. Perhaps that’s why the MSNBC Interactive acquisition is so important, since the company has indicated that it will couple EveryBlock updates with the local sections of MSNBC.com.

Have you used EveryBlock? What do you think?

12 Responses to “The Problems With EveryBlock”

  1. Tracy @ WSB

    This gets to the heart of a point I have often made to try to correct erroneous descriptions of EveryBlock, outside.in, etc. as "hyperlocal news." They aggregate information, call it hyperlocal or whatever you want to – I prefer " neighborhood." They are not news reporters or producers.

    Aside from what is put into parsable databases, somebody has to gather and publish the information that gets aggregated. THAT is where your hyperlocal news comes from – services like ours and hundreds of others nationwide. Someone else put this quite eloquently recently – this is information that cannot be harvested.

    One part of our coverage involves more than a dozen neighborhood-group meetings monthly – we are often the only journalists present – even if these groups put their minutes online (many don't even have websites yet!) eventually, that's not the same as a same-night or next-morning objective report (and often, there IS fairly time-sensitive info at these meetings).

    What still annoys me a bit (not hugely, since we worked out our revenue in another way) is that most of the support and grant money and so on seems to continue to flow to the tech side, not the content side.

    Those of us who are working the content side out on the street level, in terms of developing best practices for a neighborhood-level news ecosystem, are passed over. Just because we're not producing something that will be mirror-replicated in town after town after town (much as some are TRYING to corporatize/templatize/synthesize neighborhood news) doesn't mean we don't deserve support too. (At least we are now part of a Knight/J-Lab grant project involving some experimental editorial collaboration with the Seattle Times … the few hundred a month that Knight/J-Lab is providing for this for a year will pay for us to make a couple more freelance reporter/photographer assignments.)

    Sorry, off the soapbox. I think EveryBlock is fascinating, although it hasn't helped us in the task of producing neighborhood news – I need to look at all those databases from a more macro perspective than one zip code, one street (we cover five zip codes and 60,000 people) – so I still look at them directly myself, the semi-old-fashioned way, to see what's been filed and whether it seems to be big enough to write about.

  2. Paul Baron

    Right on target regarding nothing replaces the "feet on the street" and the local stakeholder who lives, breathes, and is engaged with the community, its people, businesses, services, and events … how is that delivered at the hyper/micro/local level — http://www.hometowntimes.com has an answser.

  3. Again, congrats to Adrian and the EB team for such great success.

    They have proven the value of aggregation — particularly in mining the municipal data that is so hard to find, but so valuable. Tracking what is happening in your neighborhood might not be a primary destination yet for you, but it should absolutely be a required destination.

    At Outside.in we have focused first on the aggregation of local media sources (currently more than tens of thousands of local media feeds daily) rather than the data because we think it can provide content and context.

    Check out our page on Capitol Hill – http://bit.ly/caphillsea

    As the data becomes more standardized and easily accessible we think that the data plus all of the other sources can provide a really comprehensive local and personal media experience.

    – Mark

  4. Viewsagent.com

    This is the beginning of an evolutionary process. Everyblock is like a 15th mariner leaving Europe in search of "the western route to India". Getting it right, financially maintaining itself during this evolutionary process, will require a long view. Its a blind man walking onto a new street and having to grope around to get even basic information. It's a start. It will require a long view…

  5. I've been a huge fan of Everyblock (in the same Seattle location… the Madrona neighborhood, specifically) ever since they launched with Seattle support. I find the permits & police activity updates fantastic (and mildly addictive), since it's information that had previously been hidden to me. Of course, your pointing out that I'm probably STILL missing a lot of information. But, I've always been so pleasantly surprised with the new information, that I've just assumed that the performance/completeness will continue to improve (and the Everyblock team has always delivered on that).

    In fact, the things I find LEAST useful in Everyblock are the Yelp and other ratings; I'd prefer to exclude them, oftentimes, so that I could focus on the somewhat impossible-to-find/unique nuggets that they're uncovering.

    I'm really looking forward to the continued development — less so as a part of MSNBC, and moreso just knowing that the team has the medium- to long-term financial support to make their vision a reality.

  6. James West

    I think the concept of a "primary news source" is a quaint, 20th-century tradition. If Adrian, Google, MSNBC or anyone could find a way to turn that concept into a realistic option, they deserve far more than merely grant money.

  7. Eric Moritz

    I've dug through the mass of Everyblock's code. I think the biggest flaw of the their approach is that they're Hyper-local at a national scale.

    There's nothing stopping Everyblock from using the blogs that mentioned the news about the bar. Since in most markets, the Everyblock developers are outsiders looking in, they pick the most obvious datasources: Local news, Police records, etc.

    The best thing that Everyblock could do is publish some tutorials on how to develop data scrapers so that external developers can submit data scrapers to the project. The code is out there, and it's not very hard to create a scraper for EveryBlock. They just have to encourage community support.

  8. Thanks for the critique, Joseph. This could be the best critical piece of EveryBlock I've seen.

    Yeah, I wasn't joking in that blog post about the site being 5% of its potential. We're looking forward to implementing the other 95%. :-)

    Adrian @ EveryBlock