Field Notes Inside an Integrated Communications Agency

In addition to the great speakers at the Omniture Summit, some of the breakout sessions teach me things about Sitecatalyst and web analytics in general that I hadn't considered previously. 

This year was no different. I took some notes on a few of the sessions and what I learned. First up was the Advanced Power Strategies.

SiteCatalyst : Advanced Power Strategies - hosted by Ben Gaines

Quick note on Ben - I've been communicating with him for years through Twitter as he's helped me with numerous Omniture issues and this was the first time I met him in person. Really good guy and always helpful. Loves his job, its obvious. His session was one of the best I've seen in 3 years of going to this Summit. Here are a few he uncovered during the breakout session:

1) Measuring engagement - while the topic of measuring engagement is controversial (another post on this later in the week), Ben and crew had some tangible ways on computing an engagement score. First off, you need to decide which events on the site qualify as 'engagement' activities. Each one of those events would get a value. You would capture that value in a counter eVar (conversion variable).

Example: Say you have 4 things on your site you to show someone is engaged with your site and assign values accordingly

Email a Friend - 5

Subscribe to a newsletter - 8

Upload a video - 2 

Write a review - 4

If someone came to your site and subscribed to the newsletter and emailed a friend, their engagement score would be 13. The values need to be thought through to determine which activities are more valuable. Now that you have that engagement score, you can relate it to campaigns, keywords, anything to use as a criteria for success. Really clever way to use eVars to create the scoring.

Additionally you could classify the scores into ranges with names such as 'Low Engagement', 'High Engagement', etc.

You could also use an incrementor event to score it so you could create calculated metrics, such as enagement per visit, etc.

Check out the Omniture blog  for a ton more on this. 

2) Use a new plugin called GetPercentPageViewed to figure out if visitors are scrolling vertically to see content. ClickMap is useful to a point but doesn't tell you how many visitors scroll to the bottom of the page to see everything. You can capture the percentage as an eVar and then classify them into ranges to see what percentage makes it all the way down. Very useful in designing pages and to understand if your content is even visible to visitors.

Again, Ben can talk more eloquently on this on the Omniture blog.  

3) Participation is a different way of attributing credit to web site activities. Typically, credit for pages is split it up amongst the pages encountered. Participation gives full credit to a page or activity so that you can figure out which content is the most valuable in driving conversion.

Example - say you have 10 pages involved with a $100 order. In the normal way attribution occurs for a page, each page would have a value of $10 (100/10).

With Participation you would give $100 to each of the 10 pages involved. That way you can determine the relative value of pages in generating conversions (or revenue). You can do similar things with other events such as a lead or subscribing for an email to understand which parts of your sites are a part of the conversion.

I've been using this one for the past few years, but wanted to share in case you weren't aware it was there.  

One of the fun aspects of every Omniture Summit is the product enhancement announcements. Much like an Apple release, web analyst geeks nervously await the new goodies to arrive with much anticipation. This year's announcements focused on 2 aspects. Both are pretty significant. 

The first is a Display Targeting Solution for Advertisers. Essentially, Omniture is allowing advertisers to leverage the data collected onsite to run offsite advertising. By using the onsite data, advertisers can create more targeted ads based on customer segments. Segments could include groups of visitors that exhibited certain behaviors such as viewing specific content or abandoning a cart or other information you know about a visitor such as source of traffic, browser type, geography, etc.

Example of how this works...say you have a group of visitors that have put Widget A in their shopping cart but for whatever reason hasn't purchased it yet. Using that knowledge, Omniture can generate targeted display ads (banners) on other websites that are geared towards closing the deal. Whether its better messaging, different offering, more relevant imagery...it can be customized to that audience. Additionally, using some of Omniture's other products like Test & Target, advertisers can do A/B and Multivariate testing to figure out which creatives generate the best results.

Or let's say you are a non-commerce site, you could leverage the data about what kinds of content a visitor group consumes, and push targeted ads that appeal to those visitors based on what they've read in the past or what other visitors have also looked at. 

Previously you could do something similar with DoubleClick's Spotlight functionality but that was always somewhat limited. This solution offers much more promise based on the richness of the data collected for segmentation.  

Segmentation, targeting and relevancy is the name of the game these days when it comes to your website and the content you are showing to visitors. Technologies like this allow for you to be a lot more precise in your marketing as opposed to carpet-bombing everyone with the same message.

The second major integration Omniture announced is with Facebook. Even better, that integration is free. By leveraging an API connection, some of the interesting data Facebook is collecting can be brought into Omniture and combined with the data you are collecting about your site.

What do I mean by interesting data? Facebook has a ton of demographic information such as gender, age, interests, groups, etc that can be brought into Omniture Sitecatalyst and used for drilling down into the data. You can find out what age groups are checking out your content, what gender likes product A, what kinds of interests generated sales, and so on. So from a customer intelligence perspective, this is huge. Facebook has 130 million users and with that its probably the largest collection of demographic data in the world.

Not only can this data be leveraged inside of the reporting in SiteCatalyst, but also used to create segments to advertise within Facebook. So again, suppose you have a group of visitors that have checked out certain content on your website, you can re-target those visitors on Facebook with relevant ads. 

The one lingering question I have on this is whether you can see all the demographic info of all Facebook visitors that come to your site or only fans of your brand. Obviously, the latter becomes less useful since you might only have 200 fans of your brand. But if it is everyone, this is the holy grail of marketers based on the sheer size of Facebook users.

Just to reiterate a point I make a lot in discussions, we aren't tracking you as an individual but in aggregate. We don't know John Smith is the one on our site doing these activities, instead we see you as some long visitor id like CWXY-1323-2233jd-45jasd. So the passing of data between these disparate systems shouldn't be a cause for concern from privacy advocates. It anonymous. What we are doing is leveraging the aggregate information to create segments for targeting and relevancy.

With both of these major integrations it is obvious what Omniture is trying to achieve, a campaign optimization platform. Omniture already has tools that optimizes content and website design. This is the next logical step, taking your online (and offline) advertising and developing technology to help marketers optimize their budgets to generate the best results. Additionally, the integrations are making it easier to create custom segments based on cross-site knowledge which will ultimately allow marketers to pinpoint their marketing and improve conversion.

Once Omniture cracks the code on campaign attribution and cross-channel influences and combines with these products, they are going to be on to something extremely powerful. Essentially, Omniture could become your campaign planning/media mix tool.  

Exciting times in the world of online marketing. 

For other posts on the Omniture Summit (and more to come over the next 48 hours):

Post 1

Post 2  

So it turns out that I lied somewhat about maintaining a semi-live blog during the Omniture Summit 2010. It proved too difficult with everything going on and because the battery life of my laptop didn't really allow for it. If anyone was jonesing for information during the Summit, you could have followed my assault on Twitter. I think I had in the neighborhood of 125 tweets in a span of 2 days. Insane. 

So now I am back in Raleigh and beginning to transcribe my handwritten notes (the ones that I can actually read) into digestible pieces for the Capstrat blog. 

I was really interested to see how the Summit was going to change after the Adobe acquisition. Once the doors opened for the general session, I could see it was the same look and feel as previous events. I've always likened it to a web analytics rock concert, complete with huge monitors and loud music, and this one didn't disappoint.

Opening session saw Josh James (head of Adobe's Omniture business unit) welcome everyone to the summit. In his opening monologue, he mentioned that we are in the midst of the Industrial Revolution of Data and that this was the decade of the CMO. Successful CMO's are going to need to:

1. Create engaging experiences
2. Unify and personalize consumer interactions across channels
3. Make the leap from marketing metrics to business metrics
4. Inform other business functions and invert accountability 

Last year I attempted to run a semi-live blog of the happenings here at the Summit on my personal blog "Diary of a Madman...", and I will attempt to do so again. I am going to opt for short bursts throughout instead of my usual 6 pages of notes crammed into one post.

I got in late last night so missed yesterday's kickoff and meeting some folks but am ready to roll this morning.

Some interesting things so far...

1) I missed the last Omniture shuttle from the airport. After learning that was the last one, I went on Twitter and asked if there were any more shuttles. Omniture monitors Twitter religiously and responded to me in about 30 seconds, and within 20 minutes they sent a car to pick me up. That is incredible! Builds goodwill, but also demonstrates how important Twitter comments are to them.

People who don't get Twitter don't understand the massive power it has. All messages get amplified, and smart companies are interacting with customers to show they are listening and are a part of the conversation. The walls of vendor and customers is eroding at a rapid speed. Is your company a part of it?

2) The vendor showcase has a lot more Adobe content to it. I wonder why?

3) When I registered I was given a Poken stick. Poken is used as a virtual business card where you can trade contact info by touching Poken sticks (which sounds funny). I think this is a great idea as one of the main benefits to this conference is the networking aspect. This tactic makes that a little easier.

On to the general session. You can follow me on Twitter @hazenj as well. 

Web safe? Web safe!? Curse the Web and it's lack of typographic variety!

Imagine a Web where beautiful typography is unlimited (with out the use of Flash).  Imagine how it would differentiate a design or a company.

But every Web designer has had it beaten into their skull that they should limit the fonts in their design to Verdana, Times, and a few others. While most hate it, many have embraced the limitation and relish the challenge. Others like the continuity it provides. Either way, the reason we do this is because nothing else works.

But the answer is not so simple, or nearly as boring.

There is a CSS property called @font-face that allows you to use any font you wish. It's not universally supported in current browsers. There are also different font formats for Internet Explorer (imagine that). Harder to solve is the issue that you can't take any commercial font and put it on the Web for everyone to have.

So technical and legal limitations killed Web Typography from taking root. But things are changing. CSS3 is on the way. The ranks of good, free, unlicensed Web safe fonts are growing. New license schemas and services like TypeKit and Typotheque for font usage are emerging. Now Firefox has released a new cross platform font format.

Keep an eye on this. I sense a growing trend.
 

A recent article in my Google Reader caught my eye entitled, "Biometrics firm confirms: User counts for websites are 2-4 times too high". The post details how a company used some really fascinating techniques to determine a more accurate count of unique visitors using individual's computer mannerisms for a website rather the traditional way of using cookies. 

Apparently everyone has their own idosyncratic pattern when it comes to using a computer including pauses between keystrokes, mouse movements, etc. Kind of like our web fingerprint. Really had no idea that existed and impressed someone thought to measure that as a unique identifier.

While the technology is intriguing, I am guessing the more important issue is the assertion that everyone's web analytics platform might be over-counting unique visitors. Not sure if most people realize this but a majority of web analytics tools utilize cookies to determine whether or not you've been to a website. The problem with doing that is most folks own multiple computers or other devices that can surf the web. I have 2 computers at home and an iPhone, so if I went to the same site on all those machines I would count as 3 unique visitors. Additionally, I believe most cookies are browser specific so even if I used the same machine for surfing but used 2 different web browsers, I would be counted twice.   

In a nutshell, the count of unique visitors has always been ripe with problems. It's probably getting worse as more devices are connected to the internet (just wait until everyone's TV is connected). This technology might help fix some of that, but in lieu of that I'd say don't worry about it. It doesn't truly matter in the end whether your unique visitor count is actually accurate. I would argue against obsessing about your unique visitor counts. Instead trends are where the relevance lies. Are the numbers still directionally correct? If the unique visitor count is 1,000 as opposed to 5,000 does that matter as much as whether the trend is up or down? Doesn't the trend dictate your action plan more than the absolute numbers? 

In the past, I've used the metric 'visit's as my key traffic metric just because of the known problems with tracking unique visitors. Additionally, to me, visits lined up nicely with similar offline actions such as a phone call or a trip to a physical store. That way I could look at conversions based on user interactions/sessions and try to improve on that rate. The key point is choose a metric whether its visits or unique visitors and go with it knowing there are inherent problems with the metric.  

The only full-proof way to get a correct count of unique visitors is to have a site that mandates a login. Instead of measuring the cookies, you'd been measuring the count of userids. However, I would imagine though you'd lose more than your ability to track unique visitors if you forced visitors to create a login in order to use your site.  

Web analytics is great in that we can track a ton of web behaviors that the offline world can't, but with that ability comes incredible complexity when it comes to accuracy. You have to just let go that things aren't 100% accurate, and in some cases aren't close to accurate. This ain't finance or accounting, but if the numbers are showing the right trends you can make the right business decisions.  

This past Saturday I had the pleasure to attend and present at the first AnalyticsCamp at UNC's Kennan-Flagler Business School. I was immediately popular as Capstrat provided the coffee and breakfast (from Fetzko's) for the event. It wasn't nearly as geeky as you might think and amazing to see a couple hundred people spend their entire Saturday discussing analytics. 

The emphasis of the Analytics Camp is to bring together all the folks involved with the different silos of analytics under one roof to learn from each other and cross-pollinate with ideas from their different disciplines and experiences. My background is in web analytics, but I got more out of hearing about developments in areas outside of my expertise, which is definitely the point.

The "unconference" had an interesting communal format as attendees propose topics for presentations in the morning and everyone votes which proposals will turn into presentations. It is a somewhat difficult event to prepare for as you aren't exactly sure if your material will make it or not. Luckily mine did, more on that in a minute.

After the voting took place we all broke up into different classrooms to hear about the different topics. Most of the presentations consisted of about 30 to 40 min of slides and then more of an informal discussion with the audience asking questions. For the first session I split my time between Text Analytics & Sentiment Analysis (VoC) and Applied Advanced Analytics for Business Solutions.

Text Analytics & Sentiment Analysis (VoC) was presented by Manya Mayes from SAS and was really interesting as this was one of the fields that is outside of my normal analytics box. The technology being used to mine text and voice to understand patterns is fascinating especially when applied to improving customer service and looking for potential problems. I thought an interesting example was mining the crash data from car accidents and looking for key phrases and correlations to things like 'accerelation' and 'problem' (something Toyota is likely doing right now). With the proliferation of verbatims throughout the internet in forums, social media, review sites, and other avenues, being able to efficiently sift through that massive amount of data and see patterns will increasingly be important for companies to be on the pulse of customer sentiment and allow for quicker resolution to issues. While I was listening to the presentation I couldn't help to think of all the applications that could be enhanced with some of this knowledge such as how a company words information or content on a website, understanding problems before they become bigger issues, or even choosing the right keywords for a PPC campaign.  

I also ventured over to hear David Larson from IBM. A few years ago IBM purchased business intelligence vendor, Cognos, and its evident from Larson's presentation and even the recent TV ads that IBM views analytics as a huge growth area. I missed most of the session but one of the most important things I heard mentioned was that their is no recipe for analytics, you have to start with the business objectives and questions. All of these sophisticated tools will spit out data and information but the real trick (and the one most people are missing) is how to ask the tools to give them the data they need to answer the business question. Couldn't agree more. Before sitting down to do the analysis you have to know what you were trying to achieve. 

My presentation "Methods for tracking offline campaign through online tools" took place in the second session. This is an area that I am really exploring as Capstrat does a lot offline campaigns and trying to show some return can be more difficult than measuring online campaigns. I presented some tips I've come across in measuring them such as using indirect proxies such as keyword changes, promotional codes, vanity urls, website behavior correlations, and geo-segmentation to name a few. I got some decent interactions with the crowd and hopefully the folks that attended got something out of it.

Lunch was unbelievable, not sure where it came from but easily the best food ever at a conference. SAS totally hooked us up. I got my fill of chicken kabobs and humus with pitas. Also, it was great to see an occasional Capstrat dart being launched across the cafeteria.

After lunch, I headed up to see Adam Covati talk about 'Social Media Impact'. Huge crowd for Adam and I thought he did a great job with his presentation and the attendees were engaged in the topic. Some great points were made about making sure you have goals for your social media, otherwise measurement is meaningless. First set up your objectives whether its getting more emails for a database, growing awareness, avoiding support costs, etc. Excellent advice was given on some specific actions that could be taken and lost-cost tools that can be utilized for measuring social media impact such as using Hootsuite, ScoutLabs, Shortened URLs, Feedburner and others. Its very obvious that the biggest theme that companies are struggling with is how to measure the impact of social media as it came up countless times in multiple presentations and discussions. As a side note and plug, Adam has a new new start-up ArgyleSocial. Go check it out.

Next up was Wayne Sutton with Location-based Analytics: Measuring the Check-in. I haven't had to solve any business problems yet around the growing phenomena of applications like Foursquare and GoWalla, but there were lots of interesting discussions about the ethics and uses for the data you could collect with measuring check-ins. A lot of the conversation centered around Wayne's new startup Tri-Out and the kinds of functionality and issues they are working on to make this type of application both usable for end-users and businesses. From a marketing perspective its great to know who your customers are, when they come, where else they go but how do you reach those folks without completely freaking them out? I didn't say anything during the discussion but was entertained throughout with the smart folks in the room. I think I have a post in the works soon about the measurement and implications of all of this, but need to do some more thinking.

Last session I attended was Nathan Gilliatt's Social Media Intelligence. The conversation was free-form and I agreed with the assertion that right now, social media is sort of on its own island but will eventually be incorporated into existing fields such as public relations or marketing. As for the measurement of social media, Nathan pointed out he believes it will eventually move into the Business Intelligence space as another piece of data that can be used by an enterprise to make decisions. I'd say right now the measurement of this activity is in its infancy, but as tools get more sophisticated in mining text and sentiment (see the SAS presentation above), I can see where this will get roped in.

I hope some of the presentations I couldn't make eventually make it to the web as it was difficult sometimes to decide between 2 or 2 sessions going on at the same time. 

By the end of the day I was completely exhausted, but in a good way. Met a ton of great people. Loads of information (and food), but the discussions that took place have gotten me thinking more broadly about analytics and how some of these things fit together. I think the best thing about this unconference is the fact that we can start to build our locale community of people in this field and learn from each other. We all bring a piece of the puzzle to the conversation. By pooling together all this talent I think we can create a very interesting network to tackle the next business challenge, create the next startup, or write the next book. Really grateful to be a part of this and hope it continues. 

The previous post in this series detailed that effective Public Relations (PR) efforts -- like any campaign -- must start with a clear objective, metrics, and baselines. This post focuses on the actual measurements and techniques you can use to measure the effectiveness of your approach. 

In doing research on the topic, I came across posts from Don Bartholomew, who writes eloquently about measuring Social Media on his blog, Metrics Man. While his posts focus on Social Media, much of his thinking is relevant to Public Relations. Either way, it's really good and you should definitely check it out for ideas. 

What are the measurements you need for PR?

If you've ever tried to measure PR, you've probably come across Advertising Equivalents (AE). Unfortunately, AEs don't really tell you anything useful. I liken it to the age-old marketing metric of "Impressions" sometimes called the "eyeballs metric." The problem with this approach is that it doesn't measure outcomes and measuring whether those "eyeballs" really saw it or not is an iffy proposition.

Another common PR metric is the number of mentions. While it's not a bad idea to try to grow mentions, it again doesn't really point to an end result. You don't do PR to simply get mentioned, you do it to serve a purpose like grow sales, change perception, or avert crisis to name a few goals. Now if you can correlate the number of mentions with tangible financial gains, I'm all for measuring that way.

So after bashing those two metrics, what's my approach?

While the best strategy depends on the specific campaign objective, I'll try to outline the most common objectives and my thoughts on suitable metrics for each: 

1) Grow Sales/Leads

What to measure: 

• • Percent change in sales/leads during times of PR or after an acceptable amount of time (in order to factor in latency, especially companies with long purchase cycles). 
• • Market Share Growth - again factoring in latency
• • Conversion Rate Increases - the theory being the more PR you do the more the awareness and perception should increase and aid in the conversion funnel. 

2) Change in Perception (i.e. Awareness or Opinion)

What to measure: 

• • Web Traffic Market Share against competitors - use tools like Hitwise or Compete to see how your traffic matches up against your competition during times of PR. 
• • Keyword growth - especially branded keywords as volume should increase based on more PR, assuming the PR is successful in driving interest. You can use your own web analytics tools or Google Trends to measure growth and even contribution to conversions.
• • Web traffic increases - Focus on visitors directly accessing your site and SEO visitors to see if the needle moved during periods of PR.
• • % Change in Awareness/Perception - based on pre and post surveys. These can be targeted to specific audience segments such as customers, non-customers, former customers, etc.
• • Net Promoter Score increases - Are customers more likely to be advocates of your brand?

3) Avert Crisis/Government Interventions 

What to measure: 

• • Does your Market Share take a hit or rebound after a crisis (be sure to factor in latency)?
• • Does your stock price take a prolonged hit after a crisis?
• • Legislative voting - Does it favor your company?
• • Measure negative sentiment - based on repeated surveys over time. Does it stay stable after crisis? 

I've only scratched the surface with potential objectives, but I think you get the idea. If you have other objectives, please feel free to send them to me for further discussion.

Techniques to prove PR had anything to do with those measurements

In looking at some of the metrics above you're probably asking yourself, "OK, those are great, but I am not sure how I can actually relate the PR efforts to these results." I'll grant you some of these are more proxies than direct cause and effect. In fact, causality is likely impossible, instead we are looking for correlation. You could simply look at your baseline results and your post-PR results and make a judgement on effectiveness, but you run the risk of overlooking whether PR was a significant factor in determining the effect. In order to gain some level of confidence in these correlations, may I suggest the following approaches:

1) Statistical Analsysis To connect your efforts to specific results, you will need to conduct statistical analysis using correlation, factor analysis, and regression. Using those terms I've probably just scared off about 90% of the audience reading this post. But if you are still reading, these statistical techniques you ignored during your 2nd year of business school can help you figure out if you've indeed moved the needle in a positive direction, especially if you've adopted the idea of using baselines (as mentioned in my first post). 

By peeling back the variables (marketing campaigns, prices, economic factors, etc) and running regression you can figure out which factors had a higher level of significance in driving the behaviors. While it sounds scary, you can actually do this in Excel assuming you have called out the right variables, have decent data, and someone on hand who can help you interpret the results.

2) Isolation is one of the best ways to go about doing the statistical analysis and determining whether your PR effort was significant in driving the metrics. There are a couple of ways to use isolation:

• • Do your PR in isolated geographies and measure the effect of the experiment versus a control sample. For example, you could run a PR effort in Virginia and do nothing in North Carolina. By measuring the differences in perception for those geographies you can isolate the PR effect (assuming other variables are constant). You can do similar things with other metrics such as sales based on geographies or keyword growth in cities where you did PR.
• • Do a survey that includes equal amount of people (or a statistically valid number) who saw your PR versus people who hadn't seen it and measure each group's perception. Surveying is not free but neither is wasting money on efforts that yield no value.

3) Use Tracking Codes in online PR efforts or even offline ones that point to the web as a way to directly understand traffic and conversions based on PR efforts. Example, when you do a press release with a url, add campaign tracking codes to the resolving url to see what these visitors end up doing on your site. Same goes for Twitter or shortened urls. This is the one way where we can see true end to end results based on behavior. 

For other insights on how to track PR (and social media), some blogs that I found to be useful:

MetricsMan

KDPaine's PR Measurement Blog  

Journalistics - How to Measure PR

If there are other blogs that folks have found insightful, please post in the comments to share. 

So this concludes Part 2...the other scheduled parts will include some further drill-down into using web analytics to measure PR and some hypothetical case studies. But are there other topics you want me to explore? I'd love to hear any suggestions for further topics in this space, so please let me know! 

1. Create an objective tied to a business goal - As with any strategy or tactic, you have to define what you hope to accomplish and you have to define it in terms of things senior management thinks are important.  You can't just do PR aimlessly and expect good things to happen. This is true with any marketing campaign, web page, collateral, etc., and PR is no exception. For example, suppose you have a growing problem with negative sentiment towards your brand/company/persona. Your objective is to combat that negativity through proactive PR. It may seem like a trivial exercise, but we often find ourselves in execution mode before we've even accurately defined the problem we are trying to address. Take the time and define the objective.

2. Agree in advance on how you will measure the objective - Without this, you are going to have problems understanding if you have been successful and securing future funding to do this kind of work. You might propose web analytics to determine how often people are searching for your brand, or you might turn to surveys if you want to assess your brand's reputation. Think through what you need to measure and choose the right tool for the challenge at hand.

3. Establish a baseline - You need to have an idea of where you currently stand in order to determine if your efforts have had the desired impact. Don't solely rely on intuition or an apparent consensus among the powers that be. Start with a real number on your organization's current performance, whether it comes from surveys, sales or lead generation. Keep in mind that the bottom line for the C-suite is, well, the bottom line. Financial measures can be difficult to tie to PR efforts directly, but people will really sit up and take notice if you can find a way to do that.

4. Be sure your PR campaign strategy and resources align with your objective - This may sound obvious, but you'd be surprised how often organizations get caught up in silos, fall back on "how we've always done it," or just get lost in tactics at the expense of carrying out a comprehensive strategy. Be sure your strategy aligns with what you already know: what customers have told you they think is important, how geographic differences come into play, and what you've learned from demographics and marketing research that your company has probably spent a lot of money to accumulate. You may have to invest in surveys for questions that are new to the organization, but don't neglect the low-hanging fruit: existing customer data, web analytics, and data on competitors from tools like compete.com or Hitwise. Using the data to craft the PR strategy will help ensure that your message reaches the right audience. 

Part 2 will explore the next steps: How to Measure and What to Measure...stay tuned later in the week.