Showing posts with label data. Show all posts
Showing posts with label data. Show all posts

Sunday, December 5, 2010

When’s the Best Time to Publish Blog Posts?

This guest post is by HubSpot’s social media scientist, Dan Zarrella.

Of all the data analysis that I’ve done, day-of-week and time-of-day data has been consistently the most popular. So in preparation for my upcoming webinar, titled Science of Blogging, I decided to combine all of my existing data on timing with my new research into one master post on the subject.

The first time I looked at blog post timing was when I was analyzing retweets. I found that retweets exhibit a strong diurnal pattern, in that they’re more common during the day and less so at night. I noticed that retweet activity tended to peak around 4pm EST, suggesting that this might be the best time to tweet a blog post for maximum potential retweet reach.

When I looked at retweet activity over the days of the week, I saw that they peaked later in the work week, specifically on Friday.

Since I first published this graph, the most frequently cited piece of this research has been the idea that Friday at 4pm is the most retweetable time of the week. While your niche maybe different, this data was based on analysis of nearly 100 million retweets, so in aggregate, Friday at 4pm is indeed the most retweetable time of the week.

Moving on from retweets, I started studying Facebook sharing and discovered some things that surprised me about timing there, too.

First, while major news sites and blogs publish articles during the work week, articles that are published on Saturday and Sunday tend to be shared on Facebook more than those published during the week. Perhaps one reason for this is that (as Wired reported), more than 50% of American companies block Facebook at work.

Next, I looked at the effect that the time articles were published had on the number of times they were shared on Facebook. I found that while there is a fair amount of variation, articles published in the morning, around 9a.m. EST, tended to be shared more on Facebook than articles published at other times of the day.

Looking back at these four data points, it may seem that they’re contradictory, but thinking through them a bit more, we can see that they is not necessarily so. Both day-of-week charts tell us that we should experiment with publishing articles later in the week—on Friday and Saturday specifically.

And by publishing posts early in the day, but tweeting them later in the afternoon, we can stimulate both Facebook shares and retweets.

I recently did a survey of over 1,400 blog readers and I asked them what time-of-day they read blogs. Morning was the most popular, followed in decreasing popularity by the rest of the day. Most respondents reported reading blogs at more than one time, so this piece of data reinforces my suggestion to publish early in the morning.

The best timing advice, however, may actually be around frequency. Last week, I analyzed 1000 of the most popular blogs on the web, according to Technorati. I compared their posting frequency with the number of incoming links and visitors they had attracted (according to Yahoo and Compete).

I found that among very popular blogs, publishing multiple times per day led to a huge increase in a blog’s success. This tells us that rather than focusing one perfect day or time, we should aim to publish at many times, and on many days.

Have you experimented with post timing and tweeting? What has your experience shown about the best times of day or week to reach your readers?

Dan Zarrella is HubSpot’s social media scientist. This post contains data from his upcoming webinar The Science of Blogging, taking place on December 9th.

Related Articles

Saturday, November 20, 2010

Gnip Gets Its Hands on Twitter’s Firehose

Social media data company, Gnip has announced a partnership with Twitter that will allow them to sell up to 50% of Twitter’s data stream to anyone who can afford the price. And it’s quite a price.

According to ReadWriteWeb, “Gnip will offer 50% of all the messages posted to Twitter for $360,000 per year, or 5% of all messages for $60,000 per year. Gnip refers to these as the Halfhose and the Decahose and they’re also offering a Mentionhose which hones in on all @replies and retweets of a specific name.

The firehose stream can only be used internally, not displayed, so its primary function is to help developers test their social media related products. It’s also a source for companies who make a living by monitoring social media and the web. Those two appear to be Gnip’s target customers, but what about marketers? Surely there’s marketing data to be mined from 50% of Twitter’s entire data stream.

This brings up several issues. First of all, the price tag will make it impossible for most business to take advantage of the offered data. The ReadWriteWeb article suggests that this will simply increase the need on the black market. Let’s stop here and enjoy the concept of Twitter data on the black market — images of men in trenchcoats on dark street corners who exchange bird calls as secret signals.

Then there’s our old friend Privacy. He really shouldn’t even stick his head in here because the very nature of posting your inner thoughts to a public forum means he’s not a part of the club. However, you know how people are. The first time a law firm buys a Mentionhose in order to support their case, there’s going to be trouble.

(Side note: Private Twitter account data is not included in the streams being sold by Gnip.)

Finally, there’s the whole idea of Twitter or any social media site as a data sales agent. ReadWriteWeb quotes Steve Rubel, SVP, Director of Insights for Edelman Digital as saying;

“I think all these companies could see that there’s more money in data services than there could be for them in advertising. No one is going to have more data than the platforms themselves – for Twitter that could be a bigger money maker than advertising.”

So is this a good thing or a bad thing? Honestly, I’m not sure. I believe that Twitter has a right to sell any public data they accumulate.  Knowing that the data is for sale won’t change the way I Tweet, but that might not be the case across the board. I imagine there are quite a few people who would think twice about what they type if they realized that their words were being saved, sold and analyzed.

In the end, though, it’s not about Twitter and Gnip selling streams, it’s about what companies do with the streams when they buy them.

What do you think of Gnips “fire” hose sale? Does it worry you, or just another cog in the great social media machine?

Related Articles

Friday, November 5, 2010

Facebook’s Data Tactics: Disrupting or Disturbing?

Many people like to talk about the disruption that Facebook is creating in the business landscape. Most prominent of those is Mark Zuckerberg himself.

Funny thing is that when he talks about the new world order that is being brought about by Facebook (and I suppose there are others but that’s not his job to say anyone is as smart as he is now, is it?), he usually leaves out the dark spots.

One of those has been revealed by Google as the search giant is pulling Facebook’s ability to import contact data from Google properties (Gmail) for the simple reason that Facebook doesn’t offer reciprocity on the deal like most other sites.

Jason Kincaid of TechCrunch reports on Google’s reasoning for why they are pulling the plug on Facebook with this response from Google themselves

Google is committed to making it easy for users to get their data into and out of Google products. That is why we have a data liberation engineering team dedicated to building import and export tools for users. We are not alone. Many other sites allow users to import and export their information, including contacts, quickly and easily. But sites that do not, such as Facebook, leave users in a data dead end.

So we have decided to change our approach slightly to reflect the fact that users often aren’t aware that once they have imported their contacts into sites like Facebook they are effectively trapped. Google users will still be free to export their contacts from our products to their computers in an open, machine-readable format–and once they have done that they can then import those contacts into any service they choose. However, we will no longer allow websites to automate the import of users’ Google Contacts (via our API) unless they allow similar export to other sites.

It’s important that when we automate the transfer of contacts to another service, users have some certainty that the new service meets a baseline standard of data portability. We hope that reciprocity will be an important step towards creating a world of true data liberation–and that this move will encourage other websites to allow users to automate the export of their contacts as well.

Of course, the goal of this action and the statement that follows it is to paint Google as riding on the white horse of openness while Facebook rides the black horse of closed, unsharing and uncooperative data monger. They actually pull it off in this case.

Kincaid continues

Facebook has claimed in the past that there are sensitive issues around exporting contact information. But that hasn’t stopped it from pulling in whatever data it can. And it has also forged deals with both Hotmail and Yahoo that will let those services access its contact data. Google didn’t do a partnership with Facebook, so it doesn’t get the goods.

In other words, the battle lines are being drawn and there is enough of a window for a company like Google to paint Zuckerberg and Facebook into the corner of their disruptive business being more of a disturbing business. A business where openness is a political front facing ideal that doesn’t quite depict the behind the scenes reality of “If you don’t play with us we won’t recognize you”.

It looks like this gamesmanship is all part of the growing process of the future of the Internet as we know it. Google is now the granddaddy of the space while Facebook is the young whipper snapper coming up with little respect for anyone and doing what it deems necessary to succeed (Which is make as much money as possible not change the world like a sweaty Zuckerberg proclaims. Mark and your minions, please stop thinking that the rest of us are morons like your college ‘friends’).

This won’t necessarily impact the actual users of these services (other than those who are silly enough to just import everything from one place to another without thinking about whether all of those folks even belong there). What it will impact is just how much of the social graph Facebook will have ultimately.

You see, despite their aggressive tactics they are working hard to piss off a rather larger part of the online world by treating Google as the enemy and not a ‘frienemy’. If Google decides to take their ball and go home that will sting even the mighty Facebook and it could disrupt the disruptor even more.

Your thoughts?

Related Articles

Friday, September 24, 2010

Insights from Ogilvy 360 DI Gov 2.0 Exchange Panelists

What is Gov 2.0? What are some of the trends and opportunities in this space? What are some of the exciting break throughs? How can an organization benefit from it? Why should I care?

On Monday September 27th, Ogilvy 360 DI will be bringing together some of the preeminent thought leaders in this space to discuss these very questions. The event which will focus on how social media tools are shaping government, the 2010 elections and issue campaigns is standing room only. Nearly two hundred people have registered on Eventbrite. Seating will be on a first come first serve basis.

This week I had the honor of interviewing a couple of the panelists that will be featured during next week’s Ogilvy 360 DI Gov 2.0 Exchange. Their responses were candid and compelling; and provided a sneak peek at the conversation and themes we will tackle on Monday.

Ogilvy 360 DI: What do you see as the vision of Gov 2.0 and why is it relevant?

A: MICAH SIFRY, Co-founder and Editor, Personal Democracy Forum

sifry.jpg

At Personal Democracy Forum, we prefer the term “We-government,” the co-creating of new forms of collaboration and service that use technology, public data and the social web to address vital issues and solve public problems, that enables us to do more with less. It’s neither Right nor Left, not small government or big government, but effective do-it-ourselves-government.

This is relevant for obvious reasons. In many ways the old ideologies have run out of steam. Laissez-faire capitalism has been badly discredited by the financial meltdown (and Enron and accounting scandals before that); big government liberalism, where wise technocrats supposedly engineer good public policy on behalf of the public, has also run out steam, captured by special interest groups that block innovation.

Meanwhile we are living in a paradoxical moment of political gridlock and technological transformation. Every day that ordinary citizens watch their elected leaders struggle and mostly fail to get anything done in the face of organized blocking minorities, makes the public feel more powerless. But at the same time, every day a new tech innovation puts more power literally in our own hands. In such a moment, should we be surprised that the single biggest Twitter burst seen around any national event was not for President Obama’s State of the Union speech this past winter, but for the unveiling of the iPad, a few days later?

People are looking for new answers, and a lot of innovation is starting to emerge from the edges, where civic hackers are inventing new ways of combining public data with community engagement. That is the promise of We-government. Not e-government, where the authorities use the web to provide the public with information and services delivered from above, but where we reinvent government as a platform connecting all of us around the issues and needs that matter in our own lives most.

Ogilvy 360 DI: In your opinion, what are some of “the success stories” that exist in the Gov 2.0 space?

A: ALEXANDER HOWARD, Gov 2.0 Correspondent, O’Reilly Media

Alex Howard

Alex Howard

First, given recent assessments and what I view as a general consensus by close observers, we’re in open government’s beta period. There are definitely both risks and rewards to the use of Web 2.0 by federal agencies.

While U.S. CTO Aneesh Chopra has hailed 2010 as the year of participatory platforms, they are still relatively early in their development. As is the case everywhere open government is in beta, the question of whether crowdsourcing national challenges at Challenge.gov leads to better solutions will remain outstanding for months time to come. While Challenge.gov may be the most visible platform, the first US CTO is thinking big in terms of using technology to meet the policy goals of the Obama administration. The re-launch of the Federal Register is definitely a success story. Other case studies include the Department of Health and Human Services, where CTO Todd Park is working on making community health information as useful as weather data. Open health data from Health and Human Services is driving more than 20 new apps. As Tim O’Reilly observed in his post on NHIN Connect and open healthcare records, “there’s some fresh thinking going on here, influenced by the best practices of open standards and rapid Internet development.”

The reboot of FCC.gov and rollout of APIs and developer engagement holds some promise. The launch of the Civic Commons code-sharing initiative could be a big deal to state and city governments. The progress of bringing open government to courts bears close watching. And if you’re tracking the technology that will make government better, it’s clear that network visualization and the mining of open data sets will have major utility for fraud detection and reduction. Abroad, the growth of government 2.0 in Australia and development of open government in Britain are key examples, particularly Data.gov.uk.

If you review the Gov 2.0 activity in summer 2010, however, you can see that the steady progress in this space. As we look ahead to the next Congress and the second half of the Obama administration’s term, it’s clear that open government is a mindset as much as a technology challenge. Taking on open government challenges at California scale is hard enough; making that transition at the federal level will be even harder.

Come join us on Monday, September 27th as we continue the conversation with these two luminaries and their fellow panelists Mark Murray, deputy political director for NBC News; Ari Melber, correspondent and blogger for the Nation magazine and Politico; and Gwynne Kostin, Director at the Center for New Media and Citizen Engagement at the GSA. To learn more about the event, read Rachel Caggiano’s post and if you have not yet, RSVP on the Eventbrite

Photos courtesy of Micah Sifry @ personaldemocracy.com and Alex Howard @ oreilly.com

Related Articles

Thursday, September 16, 2010

Phone a friend or ask the audience?

I recently bought a new (old) home in Atlanta. Temperatures have been, well, hot and the AC has not stopped running all summer long. But when my energy bill tripled for last month’s usage I had to wonder if there was something I should do. So, I reached out to my social circle; but neighbors, friends and family all gave me different opinions on what was “normal” for this time of year. I just wanted to know what normal energy consumption should be and understand how my home compared. While I’m sure there’s an app for that, I opted for a simple search which brought me to Microsoft-hohm.com where I could see how my home compared with my neighbors, the state and the nation. In fact, the site suggested that I could save quite a bit of money if I took steps to make my home more energy efficient. Decision made. Add more insulation, seal the windows and then reevaluate.

My point here is that while some decisions, like buying a pair of jeans, might require a couple thumbs up from your friends on Facebook, involved decisions, major behavioral changes and shifts in opinion may require an up-close and personal encounter with aggregated data from a substantial sample of “people like you.”

Access to other people’s information through social networks and aggregated data sites have made us more inquisitive about “how we compare.” Am I normal? Am I unique? This is the next generation of keeping up with the Joneses, just a bit smarter and less impulsive.

Conceptually, I like the ideas expressed in IBM Smarter Planet site (cl) that, “The planet has grown a central nervous system…we should  instrument the world’s systems, interconnect them, make them intelligent.”
So, not only are we getting faster access to data, but that data is also being used to make faster more accurate and effective decisions. This cyclical relationship between data collection, aggregation, consumption and sharing will only expedite with time.

Advances in technology, analytics and computational modeling all in stride with  the progression of social media – the possibilities for new engagement, connection and influence seem limitless.

The takeaway for communicators here is to know what data sources will influence  your target audience to make a decision, change behavior or shift an opinion?  And furthermore, what type of decision is your audience facing – Impulsive? Moderate involvement? High involvement?

In other words, is the individual buying a pretzel at the mall (impulsive), a pair of jeans (moderate), or buying a house(high)?  And, who or what will motivate the individual? Will he be self motivated, directed, encouraged , or permitted  to change behavior (consumption, purchase, perception, opinion)? Will this influence come from a single person, a group of friends or a substantial aggregate of data from a group of people “like him?”

When it comes to decision time will your target audience phone a friend or ask the audience?

Just a few examples to explore…
Phone a Friend
Shopping goes social:
blippy.com
swipely.com
Travel Advice gets personal:
TripAdvisor Trip Friends
Ask the Audience
Personal finance:
bundle.com
Home Purchase:
zillow.com
Energy Consumption
microsoft-hohm.com
Travel
tripadvisor.com

Related Articles