Buy
Online Revenues Double. Why? A Web Analytics Story
A good problem to have is that your online revenue growth has doubled. But if you don’t know why, it can be worrying. Ryan Nokes, a customer, told us about his quest to use web analytics to understand the change. I’ll let him tell it in his own words:
Obviously, the revenue jump was significant. We downloaded almost every report that Omniture's SiteCatalyst software had to offer and combined them into a massive Excel file. From there, we proceeded to look for the causal factors in the spike.
We first analyzed the events in a typical conversion funnel, from initial site visit all the way to purchase. In all previous years, the events in the conversion funnel related to increased revenues. But such was not the case in 2008. This was an interesting find. Not only were the traditional conversion funnel elements not applicable in '08, but revenue was growing because people were buying higher priced items.
We then sought to determine the root cause of this increased revenue/unit. This led us to look for other revenue drivers and thus the analysis of customer types, specifically, the revenue patterns generated by new customers vs. loyal customers (as defined by having visited and purchased on 3 separate occasions).
Probably due to recessionary factors, new customers are purchasing less. The way they interact with the website has no bearing on the bottom line. However, loyal customers are comfortable with the site, the products, and the service and are thus ordering more.
Ryan had cracked the case: the jump in revenues was due to loyal customers ordering more when they came to the site. To continue to grow revenues the company should focus on maintaining an excellent experience for loyal customers. An area to investigate is how many new customers are converting to loyal ones, and why or why not they are doing so.
Democratizing Data: Author David Stephenson Seeking Tableau Customers to Tell Their Stories
When someone writes in a Tableau Forum that he is writing a book about "democratizing data" meaning
- make it automatically available to those who need it
- when and where they need it
- based on their roles and responsibilities,
- in forms they can use, and
- with the freedom to use it as they choose
- while simultaneously protecting security and privacy
well, we just can't help but sit up and take notice. After all, democratizing data is what Tableau is all about. And when he says he's looking for Tableau customers to tell their stories, of course we want to help.
So I was delighted when I read more about the person behind the forum post - W. David Stephenson. David is a well-known author and e-government expert. He and Vivek Kundra are co-authoring a book called "Democratizing Data." Vivek is the CTO of the District of Columbia, a big Tableau customer in the e-government space, and is up for the top IT post in the Obama administration.) David and I exchanged a little email correspondence about the future book and I thought you would be interested.
Elissa Fink (EF):David, what is motivating you and Vivek Kundra to write a book about democratizing data?
W. David Stephenson (WDS): We share a deep passion for what Vivek terms the "digital public square": using Web 2.0 tools to recreate on a virtual basis the Athenian agora, where the people would come together to debate policy and do business. We also believe that an integrated data-centric strategy that makes valuable organizational data available to all workers, not just to elites (with the exact amount and combination determined on the basis of an individual worker's role), can improve their work by giving them actionable information when and where they need it to make decisions, can cut operating costs, break down isolation between departments and functions, and encourage collaboration.
EF:Speaking for Tableau, I couldn't agree more. And why do you think democratizing data so important right now?
WDS: There's never been a more important time for democratizing data, due to the global economic collapse and the accompanying loss of faith in government and corporations. If they follow the lead of the District of Columbia, States of Rhode Island and Utah and the U.K., they will start issuing real-time structured data feeds and invite watchdog groups, politicians, the media -- and us! to hold them accountable by analyzing this data themselves, rather than simply relying on their reports and their interpretations. While governments have taken the lead in this sort of transparency, banks and other corporations have also lost their credibility, and would be well advised to consider their own transparency initiatives.
Equally important, if government agencies start creating Web 2.0 tools such as tags, topic hubs and threaded discussions, they may find that visualizations and other interpretations of the data will result in crowdsourcing, in which innovative new approaches will result from the discussion and interplay of perspectives.
Finally, democratizing data can be crucial as companies have had to lay off workers. The remaining workers have had to shoulder new responsibilities, so they need more information, more tools to help them do their jobs and collaborate with other workers to increase overall efficiency.
EF: There's definitely pressure everywhere to do more and make better decisions with less - and if democratizing data can help, enlightened companies will be signing up. So what is democratizing data? Is it more than just sharing static views of data with more people?
WDS: It's much more than just sharing static data with more people! It begins with an attitude: that organizations need to be "data centric," with structured data distributed automatically at the core of their operations, accessible to a wide range of users, from employees to regulators to suppliers -- all from that central data hub (rather than being captured by proprietary applications -- a great argument for Tableau!). The key elements are:
- 1. structure the data: that metadata will stay with the data as it is used, so it can be accessed and shared equally by an infinite number of applications and devices
- 2. syndicate the data on a real-time basis to maximize its utility: in many cases this will be the first time workers have had access to real-time data, when and where they need it
- 3. release it externally as feeds that can be analyzed by the public
- 4. if you really want to benefit, "crowdsource" with the data as the District of Columbia did with its Apps for Democracy contest: developers were invited to use one or more of the data streams to create open-source applications that would serve the public. For a total cost of $50,000 ($20,000 of which was for prizes) the District got 47 apps: an ROI of 4,000%!
EF: I'm glad to hear it's so much more than just sharing static views - being able to interact with data is key to what people learn from it. So what do you see as the major challenges to organizations who want to democratize their data?
WDS: The major obstacles are inertia: it's been so difficult to share data for so long, that now we have the technology to do so, perhaps management can't even visualize the potential. One striking example, IMHO, is that the Netherlands has launched an incredible experiment, the Dutch Taxonomy Project, which allows companies that otherwise would have to file annual reports with 30-40 different agencies & companies to instead file a single XBRL data file (which all of the agencies then access automatically!) which would both save them and the goverment large amounts of time and money. However, only a relatively small percentage of companies take advantage of the program (and, by the way, once they've gone to the time and effort to tag the data for these official filings, they could amortize the cost and increase the benefits by also distributing the data internally).
EF: Given that we recently passed “data privacy day”, how “democratic” do you think data can get? Do companies have to worry about data falling into the wrong hands?
WDS: Data security is a tremendous problem, but that can't be used as the justification for not democratizing data. A great example of how to do it right is what Dr. John Halamka and the IT staff at Beth Israel Deaconess Medical Center (BIDMC) in Boston have done with the Online Medical Records (OMR): nothing is as fraught with privacy issues as medical records, and, at the same time, as critical for those with a need to know (especially ER docs) to have real-time access to all the information. Halamka has done that, and you can't tell me that corporations can't do the same, if they begin with that data-centric perspective and then determine various levels of access depending on the person's role.
Broadband Stimulus Applications and Awards
$787 billion dollars- it is likely that you have heard that figure mentioned in the past year. This is the final price tag for "the Stimulus", or The American Recovery and Reinvestment Act of 2009. Unfortunately, the program is so monolithic and the figures involved are so large that even the most inquisitive minds can be quickly overwhelmed. Instead of trying to tackle everything at once in a visualization, we decided to dig deep into just one program: The Broadband Stimulus fund.
Totaling just less than 1% of the total stimulus, the Broadband Stimulus fund may seem small but at over $7 billion dollars it is far from a trivial program. Essentially charged with bringing high speed internet to a wider audience (and spurring employment), companies and governments were allowed to apply for grants and the most efficient applicants were chosen. In the viz above, you can see the total for each state broken down by the amount of requested grant, loan and private investment. If you click on Maryland, you can see that Hughes Network Systems has proposed several programs totaling over $2.5B, though most of that number is private investment from Hughes itself. It should be noted that the location of each dollar amount is determined by the location of the company, not of the project.
The viz below can be read in a similar fashion to the first, but it details those applicants whose projects have been approved and granted funds. It is a little surprising that so few grants have been granted to companies or governments west of the Rockies. California, for all of its tech prowess, has not even garnered a single grant thus far.
Things we like about these visualizations
- Before you even try to delve into the deeper parts of the viz, you can see the total broken down in a useful way.
- As you move down the viz, you are taken to a successive level of detail. Intuitive!
- The tool tips. Clean, easy-to-read and informative. They add to the analysis rather than distract from it.
A Conversation with Jock Mackinlay: The Two Success Stories Driving Visual Analysis
One great thing about working about Tableau is that I get to work with smart people who have years of research experience investigating how to help people see and understand data better. Recently, Jock Mackinlay, Tableau’s Director of Visual Analysis and one of the world’s most expert information visualization scientists, and I had a chance to spend some “think time” together. We talked about a recent article he wrote on the topic of collaborative visualization.
By way of background, the article Jock wrote was a technical perspective titled "Finding and Telling Stories with Data" on an important research paper regarding collaborative visualization; the paper itself is titled "Voyagers and Voyeurs: Supporting Asynchronous Collaborative Visualization" (both pieces require website registration before you can access them). Jock's technical perspective and the paper were published in the magazine "The Communications of the Association for Computing Machinery", or the CACM, which is one of premier publications for computer professionals.
Elissa Fink (EF): Jock, tell me more about this recent article you wrote.
Jock Mackinlay (JM): Well, I was asked to write an introduction to an important paper about collaborative visualization. As part of that, I wanted to talk about Visual Analysis as a powerful method for finding and telling stories with data. The idea of telling stories with data is moving from research into widespread use where collaboration is a necessity. And it’s being driven by two success stories of computing: databases and visual interfaces.
EF: What do you mean by the “success story” of databases? I’ve never really thought of databases as being either successful or unsuccessful. And frankly, sometimes when I use databases, I feel pretty unsuccessful.
JM: Well, if you think about it, before databases, there were just big, messy files. The success of databases began in 1970 with Codd’s relational model, which supported transaction processing. And before you ask, Edgar Codd was an IBM computer scientist working in Britain. His invention – the relational database - truly revolutionized the world of business. By the 1980s, you saw the widespread reorganization of businesses, hospitals, schools, and governments to use relational databases. More recently - mostly throughout the 1990’s - due to the internet and accumulation of online data, we’ve seen the development of multi-dimensional data warehouses designed to support reporting and analytical queries. And since then, the trend has been a steady increase in the number and size of databases that support analysis.
EF: O.k. You’re right – I can see what a big impact databases have had on our world. But how do they and visual interfaces relate to story-telling?
JM: First, don’t forget about the success of visual interfaces themselves. They began in the 1960s with graphical user interfaces (GUIs), which replaced command line interfaces by exploiting the power of the human visual/motor system. Then in the mid-1980s, advances in computer graphics hardware prompted research on visualization, the use of interactive, visual representations of data to help speed and amplify human understanding. The early focus of visualization research was on individual analysts trying to find stories with data, first in the area of scientific data and then more generally with abstract information.
EF: That makes sense. I, as an analyst, often work by myself trying to find what the data is saying. If that was the early focus, what’s changed now?
JM: Since 2000, the research focus has expanded. Now we’re going from the visualization of an individual analyst to visual analysis – in other words, the use of visualization in larger processes of “sensemaking”. Reduced to its essence, visual analysis has a four-part cycle. First is the focus on a data-oriented task. Second is the need to forage for relevant data. Third, you visualize the data. And then you perform an appropriate action in response. Given a task, analysts forage for relevant data, which is mapped to visualizations that exploit the power of the human visual system. Visualizations lead to findings, which prompt actions. When the actions are new data analysis tasks, the cycle repeats. There are also internal cycles in this problem-solving process. For example, visualizations can indicate the need to forage for new data.
EF: I see that. Before I ever used Tableau, I would get a file of some data, do some analysis and then realize I needed more or different data to complete my analysis. It was frustrating.
JM: Yes. Now you see why commercial adoption of visual analysis depends critically on easy and efficient connection of visualization technology to databases. Thus we’ve linked these two success stories of computing. Chris Stolte’s Polaris system from Stanford (which was described in an earlier CACM research highlight) made a lot of progress on this topic. It formally specifies data views that compile into database queries.
EF: And of course Polaris was the origin of VizQL, which is at the heart of Tableau. But how does linking databases with GUIs relate to collaborative visualization?
JM: Well, my thesis is that the speed of adoption of visual analysis will be determined by a second research topic: collaborative visualization, in which telling stories with data plays a central role. Extending visual analysis from an individual to a collaborative activity increases the visibility and value of the technique. Most analysts must collaborate with colleagues and managers before actions are approved. Data views make data understandable, which encourages collaboration with people who are not skilled analysts. Interactive data views allow people to do their own analysis with data views authored by others.
EF: Yes, pretty much every analysis has some audience or set of participants.
JM: The idea of collaborative visualization is the subject of the research highlight, "Voyagers and Voyeurs: Supporting Asynchronous Collaborative Visualization" by Jeffery Heer, Fernanda B. Viégas, and Martin Wattenberg. The authors describe a prototype web application that includes several techniques for supporting collaborative visualization, and report on user studies involving the prototype. The most interesting aspect of the prototype is a bookmarking mechanism that supports doubly-linked discussions.
EF: Explain that please?
JM: Data views have the property that the same view can be specified in multiple ways. In this paper, the authors describe how to associate a bookmark with a data view rather than the various specifications of the data view, which supports asynchronous discussions about views. The most interesting aspect of the user studies was that their subjects switched between data-driven exploration and social navigation – that is, between being data voyagers and data voyeurs.
EF: I like that distinction: voyagers and voyeurs. As a business person, you find yourself in either role often.
JM: The "Voyagers and Voyeurs" paper represents an important early step in research on collaborative visualization. The authors made the excellent choice to focus their prototype on the US census data set. By focusing on a single public domain data set, they reduced their prototype’s data foraging complexity, thus encouraging their users to focus on collaborative activities. Furthermore, many people are interested in census data.
EF: So if this paper is an important early step, what’s next?
JM: The next step for research on collaborative visualization is to address topics that arise in more fully featured visual analysis applications. For example, most visual analysis tasks involve multiple data sources. Unlike the US census data set, many data sources have security issues. That can make collaboration more difficult. And, some data sources also change rapidly, making asynchronous conversations more complex. Finally, tasks involving multiple data sources often require conversations that compare and contrast multiple data views that must be viewed simultaneously.
EF: Collaboration clearly adds a lot of complexity to any analysis. But it’s critical – everyone works with someone.
JM: Indeed. And by using collaborative visualization, people can work more quickly and appropriately together. Finding and telling stories with data can help people understand the world more clearly. For example, the mortgage-backed security crisis might have been averted if mortgage data had been available for storytelling. The key is to have visual analysis technology and use it appropriately.
Suggested reading:
S. K. Card, J. D. Mackinlay, and B. Shneiderman. Readings in Information Visualization: Using Vision to Think. Morgan Kauffman, 1999.
D. M. Russell, M. J. Stefik, P. Pirolli, and S. K. Card. The Cost Structure of Sensemaking. In Proc. of ACM CHI’93, 269-276.
C. Stolte, D. Tang, and P. Hanrahan. Polaris: A System for Query, Analysis and Visualization of Multi-dimensional Databases. Communications of the ACM 51(11), 75-84, 2008.
The Viz Police: Overhaul of Recovery.gov
We at Tableau applaud the Obama administration’s efforts to improve accountability and transparency. The best way to kill germs, as they say, is to put them in the sun. We’re big believers in a digital democracy: making information accessible and understandable to the people who make decisions by it. In a democracy, that’s voters. In a company, that may be a workgroup or division.
But how do I say this nicely… the representations of data on recovery.gov are uninspired at best. In this post the Viz Police offer a little help.
Floating Bubbles. Really?
The first chart on Recovery.gov, the site that explains the new stimulus bill, is a bubble chart. Bubble charts can be useful at times but humans don’t tend to compare areas as well as lines. We have three beefs with Obama’s bubbles:
- Having the bubbles floating in space makes it hard to visually compare amounts.
- The colors don’t encode any additional data.
- The bubble labels don’t match up with the actual line items: you’d be hard-pressed to find “Protecting the Vulnerable” in the text of the Stimulus Bill, which makes it hard to dig deeper.
The graphic is pretty, but if you’re trying to understand where the money is really going, not too useful.
In the public interest, we re-created the chart. We’ve added jobs to the view but not broken it down by cost per job because, as Stimulus.org notes, without a time dimension it’s impossible to get a meaningful metric. Using a tool like Tableau you could drill down into each program to explore the data further.
Which brings us to our second point:
Put the Data Online
You can get the text of the stimulus bill online in a document, but you can't get the data by line item in a chart. That leaves people with the choice of digging through the bill line-by-line or accepting the summary-level graphs on recovery.org. We worked with a datasheet from Google Spreadsheets that was drawn from Stimuluswatch.org, which asked people to vote on projects. As a result, our dataset is not the final one in the bill.
President Obama: You won the presidency by assuming Americans are smarter than we’re given credit for. Put the all the data online so we can discuss and bring new thoughts to the debate.
Make the Data Visual
Finally we look at the other visual representation featured on recovery.gov-- a map of where the dollars will go. In this case we’re shown jobs per state, but we have to hover over each state to get the data. Again, it makes it difficult to compare data state by state. And jobs, cost and program data are split up, perhaps intentionally. But we can get a lot more out of this data by looking at them together without even creating any bogus measures.
Perhaps we want to see how the jobs split up by state. In this case, displaying bubbles on a map helps us visually compare job creation in terms of a known reference point—the map of the US. And since we often think of national bills in terms of region, I’ve added color by region. Now it becomes obvious that the South and the Rust Belt parts of the MidWest should do well in terms of jobs.
Now let’s look at what programs are going on in each region.
The Digital Community to the Rescue
Other sites have done better. The aforementioned Stimuluswatch.org has done a nice job of fostering discussion on the bill, and made its data set available. OFF the MAP did a great visual analysis of the US Economic Stimulus and Unemployment by County.
Despite our criticism of the visual presentation of this data, we’re ecstatic that the Obama administration is taking steps to make more data available. And we’re here to help make that data comprehensible to human beings, who must ultimately make decisions based upon it.
Tableau Goes to College with Institutional Research
Over the last year, we at Tableau have noticed an interesting group of early adopters of our visualization software: Institutional Researchers at universities and colleges. You probably never heard of these folks while you were in school, but they’re the ones who figure out everything from the average SAT score of a school to when to schedule classes.
Bill Tobin, head of Institutional Research at DePauw University, has long been doing interesting data analysis with Tableau. For example, want to know when the peak times for Arts & Humanities courses are? Between 10 am and noon. Math majors get out of bed earlier; the peak time for Math & Sciences classes is just before 10 am.
Here's one of Bill's visualizations:
Bill also uses Tableau to create a “truly strategic planning document” for DePauw and to cut down on the time he had spent reporting on student data.
One of the characteristics of people who work in universities is that they tend to like to share information, and we’re hoping others will use Bill’s work as a launching point for their own. For that reason, we’ve written up some of Bill’s work in a case study. He’s also done a webinar on Institutional Research.
We're glad Bill and Depauw find Tableau useful in their mission. Go Tigers!
What’s the Price of Love?
If you want to advertise on Valentines’ Day, what’ll it cost you? We entered 20 terms into Google Adwords’ Traffic Estimator and got some cost data. Then we put the terms into groups describing Love, Friendship, Dating, Gifts, and Breakups. Then we graphed the data. Yes, there is something wrong with us. And please forgive for the pink and red colors throughout—at least we didn’t use little heart icons everywhere.
Dating and Gifts will Cost You
Keywords relating to Dating are the most expensive to buy, and "online dating" is the most expensive of the bunch. That’s no surprise: Match.com, eharmony and their peers are all trying to lure you to their site. And Gifts are expensive, too. Higher cost-per-click (CPC) figures mean that advertisers have figured out how to make money buying these terms. Searching for "flowers" on Valentines’ Day? Chances are good you’ll buy something.
You Can Buy Love, but It’s Not an Effective Strategy
Terms related to love (passion, love, marriage, crush, and romance) get high search volume. But they’re cheap, from a CPC perspective. That’s probably because advertisers haven’t found a good way to make money on the squishy, goofy feeling of love itself—they only make money on the pursuit of love through dating and gifts.
Can we find the value of passion, love and dating through Adwords? Perhaps not. But it does provide some nice eye candy on Valentines’ Day.
What Should We Eat Today?
Getting Our Bearings in our New Silicon Valley Location
As you may know, Tableau Software is headquartered in the Fremont neighborhood of Seattle, WA. It’s surrounded by unique restaurants and local flavor. What you may not know is that Tableau also has a California office in the Bay Area. I moved from our headquarters to the California Office, which we affectionately call Tableau South, about a year and a half ago as number four in our small group. But we’ve grown quite a bit since then and I am pleased to announce that we’ve recently moved to a new space in downtown San Mateo, CA. The move has been exciting for a number of reasons including natural light, lots of room, fully duplexed phones, a more cohesive office space, a kitchen, and the list goes on. But the thing I think we are most excited about is the location. Specifically, we are thrilled by our proximity to a cornucopia of new lunch spots. Our upgrade has taken us from a rotation of five mediocre establishments that often required driving to an endless bounty of delicious eateries-- all within walking distance! In fact, according to the San Mateo Downtown Restaurant Guide there are 143 eating and drinking options just waiting for us to experience.
Keeping Track
After our first week, it became clear that we were going to need a way to keep track of the places we’ve tried and whether we’d go back. We could always print out the restaurant guide and annotate the hard copy but seriously…we’re way too geeky for that. Below are some views from the workbook we now have published to Tableau Server so everyone can log on and answer the age old question: what should we eat today?
This first view is just a map of all the restaurants colored by whether we’ve tried it. As you can see, we have our work cut out for us. But this map also includes coffee shops and dessert only places. Although our resident triathlete could probably have dessert for lunch, the rest of us need real food.
The view above is static but on Tableau Server there are all kinds of interactions. For instance, you can select a restaurant and link to it on Yelp to see its menu, reviews, etc. You can also simply hover over a restaurant to get its exact address and phone number. And of course you can zoom, filter, pan, and all the other usual stuff.
Other Fun Stuff
Here’s another fun view that shows the distribution of cuisines and the percentage that we’ve tried. When we first moved in there seemed to be an unnatural number of Pizza places around town. After looking at this view, it looks like San Mateo is actually overflowing with Japanese & Sushi restaurants. We’ve only tapped 10% of the Japanese & Sushi opportunities!
So I know what you’re wondering – where are the ratings? Well yeah, we’re keeping track of our own personal ratings as well as specific days we visit and average individual pricing. But all that can be found on Yelp or City Search until we get more data.
Let me know if you have any recommendations for a lunch place in San Mateo. Otherwise we’ll continue to work our way through the list. Feel free to download the packaged workbook if you want to play around with the hyperlinks or just browse the dining options.
SEO Keyword Analysis: How to Use Data Visualization to Make it Quick and Easy
Do-it-yourself keyword strategy for under $10.
The foundation of a good search engine optimization (SEO) strategy is a well thought-out keyword strategy. Unlike the instant gratification of pay-per-click advertising, organic search efforts often take six months or more to show results. During those six months, your business will be creating mountains of content and building targeted links, all focused on 5-10 specific keyword phrases. If you've targeted the wrong phrase – perhaps it was too competitive, or there wasn't enough traffic, or the traffic didn't convert – the wrong keyword strategy will have cost your business thousands of hours in wasted effort.
A thorough analysis can easily take several days and require hundreds of dollars worth of analysis tools. Hiring a specialist who knows how to use them costs even more. Here's how to use data visualization to get it done in a few hours and for just a couple bucks.
Get the Tools
A wide range of factors come into play when evaluating keywords' potential for SEO. Of these, the most important are search volume, conversion rates, competitiveness, and relevancy to your site. With endless rows and columns of keyword data, the intuition required to balance these factors is a blend of art and skill that can only be developed with time.
However, data visualization allows us to do the same thing cheaply. The human brain can process pictures much more effectively than tables of numbers. By incorporating different parameters into different dimensions of our visualization, we can quickly scan a huge list of keywords and pick out the ones worth pursuing. In this guide, we'll use a number of free tools to collect the necessary data, build visualizations, deliver actionable insights, and set benchmarks.
- Microsoft adCenter Add-in Beta for Excel (free)
- Google Adwords Keyword Tool (free)
- Google Adwords Traffic Estimator (free)
- SEO Chat Keyword Difficulty Check (free)
- Wordze 1-Day Trial ($7.95, optional)
- SEO for Firefox (free)
- SEObook rank checker (free)
- Tableau Desktop Trial (free)
Get the Lay of the Land
I will be helping Joe, owner of imaginarycreaturesfanclub.com, develop a keyword strategy for organic search. (An imaginary website about imaginary creatures – how appropriate!) At this point, I know nothing about the competitive space.
I begin by getting a sense of what broad keyword categories are worth competing in. I just want a high level overview for now – we'll do a deep dive a bit later.
I ask Joe to give me a list of the main websites in his space and add them to an Excel spreadsheet. I do a few searches of my own and add the top results to the spreadsheet as well. The Keyword Extractor from Microsoft's Ad Intelligence Add-In now saves me hours of effort. It scrapes the (potentially hundreds of) webpages in my spreadsheet and tells me what words they're optimized for. Why go through all the bother of figuring out the best keyword categories when my competitors have already done it for me?
The Keyword Extractor tool gives me hundreds of keyword ideas, some of them relevant, some of them not. But looking through the list, three main categories jump out at me: "bigfoot", "dragons", and "unicorns". I divide my keyword ideas into these three categories and throw out the ones that don't fit.
Keyword Data Points #1 and #2: Search Volume and Value
Now it's on to Google Adwords! I take my bigfoot keywords and feed them into the Adwords Keyword Tool, returning a huge list of keywords, all related to bigfoot. I do the same for dragons and unicorns and then feed the whole list into the Google Adwords Traffic Estimator. Traffic Estimator tells me the estimated cost per day for each keyword, which I use as a rough approximation for its value. Because advertisers pay more for keywords that convert well, this allows me to recognize keywords that may be particularly valuable, even despite lower search volumes.
Keyword Data Point #3: Competitiveness
Next, it's off to Wordze for a $7.95 one-day trial, where I can feed the entire list of keywords into Wordze and get a Keyword Effectiveness Index (KEI) for each phrase. You're welcome to use your favorite tool here – I'm just looking for a rough measure of each keyword's competitiveness and performance. I picked Wordze because it's cheap and because it lets me cut and paste my keywords in bulk.
KEI attempts to weigh the competitiveness of a keyword against its traffic potential, so it aggregates a lot of the numbers we've already collected. For many people, that's a good thing – choosing the right keyword strategy requires balancing so many factors that an arbitrary weighting and aggregation of the factors is sometimes the only way to reach a decision.
However, we'll be using Tableau Desktop, a data visualization package that will allow us to easily incorporate several dimensions into our analysis. So for us aggregation is bad – we would prefer transparency so that we could use the underlying data and do the analysis ourselves. But that's OK, we're just using KEI as a quick first-round screen to make sure we don't miss any high-potential keywords.
My first goal is to pare down this list of several hundred keywords. Here's a screenshot of my data in Excel. I want keyword phrases that have a balance of high traffic, high value, low competition, and high relevance to Joe's website. How am I supposed to find the keywords with the highest potential in all this data?
Narrow the List of Potential Keywords
Fortunately, this is where Tableau excels. In five or six clicks, I built this visualization of all the terms in my spreadsheet. I've separated the keywords into their umbrella categories (bigfoot, dragons, unicorns) and within each category, ordered the keywords by Estimated Cost / Day. The length of the bar corresponds to the Avg Monthly Search Volume (I'm using a logarithmic scale here, or else the highest trafficked words simply overwhelm the visualization), while the color is determined by the phrase's KEI.
Now I don't have to strain my eyes struggling to make sense of hundreds of rows and columns of data. It only takes me a minute or two to scroll through my visualization and pick out the most valuable keywords (higher in the list) that have either high traffic (long bars) or good KEI's (green color).
Additionally, look at all the green for the bigfoot category vs. the unicorn category (see full keyword visualization). I've already got a preliminary indication that I probably want to focus my efforts on bigfoot instead of unicorns. That's an actionable insight that I never would have noticed with a table of numbers.
Make the Final Keyword Selections
I was using KEI as a rough measure of keyword competitiveness, mainly because there's a cheap tool to allow me to compute KEI for phrases in bulk. Now that I've narrowed my list of potential keywords, I can use a tool like SEOchat's Keyword Difficulty Check to evaluate the competitiveness of each term by hand.
I copy my last visualization and replace the KEI measure with my new Difficulty measure. I tell Tableau to ignore the keywords for which I didn't collect a difficulty score, and almost instantly I have a new visualization.
Just looking at the bigfoot terms for now, it's clear that "bigfoot sightings" is the dominant phrase in search volume. That dark red color also tells me the term is extremely competitive – Joe won't be ranking for it anytime soon. For now, let's hide the phrase "bigfoot sightings". Tableau rescales my graph and new insights jump out.
"Bigfoot sightings" was too competitive for my liking, but there are some variants that look very promising. "Sighting of bigfoot", "sightings of bigfoot", "first bigfoot sighting", "new bigfoot sighting", "Oklahoma bigfoot sighting", and "recent bigfoot sighting" are all very nice shades of green. These phrases are closely related, so it will be easy to create plenty of quality content to target all of these phrases. And the best part is that all that effort will be building up relevant authority for Joe's eventual run on the big-money phrase, "bigfoot sighting".
Oh, and there's even a misspelling that looking interesting. "Bigfoot sitings" doesn't look competitive at all – one highly optimized page might net me a few hundred extra visitors a month.
"Bigfoot organization" and "bigfoot photo" also stand out as highly trafficked, very valuable, but less competitive phrases. I make a mental note to evaluate them more carefully later. When I get a chance, I use SEO for Firefox to examine the competitive landscape. The top few results on Google have fairly high pagerank and a lot of incoming links, but the results from #5-#10 are unimpressive.
Joe is an avid blogger, has a very active forum on his site, and hasn't done any SEO at all until now. I think he could be a contender in search – we'll keep both those phrases in mind.
And just like that, we have a keyword strategy for the "bigfoot" category. We've got a nice set of niche keywords to attack for some immediate traffic, two somewhat more competitive terms to shoot for in the medium term, and one dream phrase ("bigfoot sightings") that we'll be building SEO karma for with all our groundwork right now. I can repeat this process in just a few minutes for "dragons" and "unicorns", and we're almost good to go.
Set Metrics
There's one last thing, though – benchmarking. Where are we now vs. where do we want to be? SEObook's Rank Checker will give me my current rank for a long list of keywords. In seconds, I can export the data to Excel and add it to my current visualization.
For visualization, I prefer to look at 1/rank rather than rank itself. First of all, it means that longer bars are better, which corresponds with my natural intuition. Second, it magnifies the difference between, say, a #3 and a #4 ranking term vs. a #103 and a #104 term – which is as it should be.
Turning to the unicorn category this time, it's clear that Joe's current SEO strategy (or lack thereof) isn't working for him. He's ranking really well for terms with no traffic at all and is barely on the board for higher volume terms. This visualization will be a great benchmark to refer to in six months when Joe wants to know whether our work has paid off.
It's up to you to decide how much time and money you want to invest into developing your keyword strategy. If your budget permits, by all means hire a specialist. But data visualization allows even non-technical users to evaluate keywords across a range of different factors. By representing different factors on multiple dimensions of our visualization, we can scan a huge list of keywords and quickly pick out the most promising.
DC's CTO to Take Top IT Post in Obama Administration - Big Believer in Data Openness
Nextgov.com and other sources are reporting that Vivek Kundra, the District of Columbia's chief technology officer to take the top information technology post in the federal government.
This is good news for those of us who believe that analytics and business intelligence need to go way beyond their current audiences. Kundra already has experience in bringing analytics and BI to the masses. Since arriving in DC 2 years ago, he's made virtually every database public. This is the real meaning of a digital democracy: when citizens are informed, they can think through lawmakers’ proposals and comment in an intelligent way. And it’s a hot topic right now.
The City of DC is a key Tableau customer - we're proud to work with them. We've got some examples up on our e-Government applications pages. Their data program manager Phil Heinrich spoke not too long ago with Ted Cuzzillo of TDWI about the City of DC's analytic efforts using Tableau. District of Columbia's Capstat Presentation Using Tableau
Obama’s administration has already shown a predisposition to openness. We'll definitely enjoy seeing how Kundra brings his brand of openness experience making that real to the Federal level.
