TruthyPR

Four must-do analytics tasks for every nonprofit to start 2010 right

So often I come across nonprofits and trade associations that are interested in analytics but don't know where to begin.  I've listed here a few small must-do things that you  can tackle at the beginning of the year to lay the foundation for becoming an analytics-driven organization in 2010.

Being an analytics-driven organization means you'll be getting better results with the same resources.  Who doesn't want that?

Establish Key Performance Indicators that your management understands

You need at least one identifiable goal that doesn't have to be explained in techspeak to your management and is an undisputed benefit to the organization.  Here are some good ones:

1. Number of donations to the website or number of donations to the website per 1,000 pageviews or visitors.
2. Or if you're an advocacy organization: Number of letters to Congress on an issue in key districts.  Note that if you feel comfortable with the geo-segmentation Google Analytics does, you can actually compute a conversion rate based upon geo-traffic, and make this "number of letters generated from web visitors from key legislative districts".
3. If your organization recruits people for a campaign, like the Business Software Alliance's whistleblowers campaign, you could use the number of incoming web leads or the rate of those per 1,000 pageviews or visitors.  (Disclosure: The Business Software Alliance is a Virilion client)

And here are some bad Key Performance Indicators.  Most of them fail one of my colleague Shayna Englin's SMART tests, because they aren't: strategic, measurable, ambitious, realistic, or time-bound.

1. More email addresses: you can have a 1 million address email list, and still not raise a dime from it.  This is not strategic, nor is "more email addresses" a measurable goal.   How many is "more"?
2. Double our unique visitors in the next year:  Very few organizations have mere traffic as a goal.  The Kaiser Family Foundation may be one example, but even then, there's plenty of traffic that has no benefit to your organization.  You need to segment out your visitors more and give them specific goals.

Configure your web analytics product to measure that goal

Most of the work of 2010 should be about figuring out what contributes to your success of your Key Performance Indicator.  Is my success newscycle driven?  Do endorsing blogs play a key role?  Do my supporters come in droves because of a specific part of my mission?  Do my visitors who look at my "About page" give more than those that don't?

You need to setup your web analytics product to measure that goal, because the rest of the year will be spent learning what works and what doesn't, analyzing the data trails of people who did what you asked.

You'll need to give reports to your management showing what works, and most importantly, what doesn't.  "Wait, Shabbir, you just told me to give my management a report showing that I'm doing something that doesn't work.  Isn't that bad for my career?"

Here's the thing, if you're doing it, and it doesn't contribute to your goal, you probably know it already, and you continue doing it because your bosses haven't seen that fact.  If you make a point of demonstrating what parts of your work are effective and which ones aren't, that's the first step to getting your managers to make your job less frustrating.

(If you work at a place where you nobody cares about effectiveness, then you'll have to look a little harder for a champion.)

Apply for a Google Grant

A Google Nonprofit Adwords grant is the most under-utilized tool for bringing in supporters I've ever seen.  Most nonprofits spend no more than $1k-$2k per month of their $10,000 maximum spend using the default ad campaign the Google Grant team setup for them.  

If you don't have a Google Grant, you need to apply for one right away.  Approval won't happen overnight, and the sooner you apply the sooner you can start generating good traffic with it.

Start measuring your rank on core keywords

Odds are good there are search keywords out there that you want to rank high on organically.  After all, if you do microloans, and Google searches on "microloans" and "microcredit" show your website 7th, six other organizations are enjoying your donors for lunch.

Take a moment and look at your website (and your mission) and come up with some core keywords.  Then use a tool such as KeywordSpy to look at the keywords of competing organizations.  Set your list for the year and then use a free tool such as RankChecker to monitor your organic search rank on these terms.  Over time you will find opportunities to spot trends before your competitors and then build out a set of useful pages that rank well for both existing and new keywords.

Actionable Analytics: Your "Visits" graph doesn't mean anything, look deeper.

Your total website visits graph is an unactionable metric

Have you ever had this experience talking to the person who does your website's analytics?  (I've not only had this experience, I used to deliver this reporting fallacy).

"What you have here is your visits graph, showing that visits are holding steady (or rising, or falling, or whatever)."

Unless you've got a number of traffic-driving campaigns running, this graph doesn't mean anything.  It's a combination of actionable analytics rolled up into one big un-actionable one.  I had my own moment of epiphany about this when I presented this graph to a favorite client who had no blog outreach, no managed advertising, and no active SEO campaigns going.   I showed her the graph of visits and she said, "Is that good?"

The answer is it doesn't matter.  If you don't have a plan to generate visitor traffic, checking the visits you've got is even less meaningful than checking for rain.  So let's say you've seen the light and want to turn this into an Actionable metric, what do you do?

Step 1: Look deeper at your Search traffic

The first step should be to look at the components that make up your visits.  Most analytics products break these up into "Direct Traffic", "Referral Traffic", and "Search Engine Traffic".  Search Engine traffic is a great example of this.  Your search engine traffic is probably one fourth to one third of your traffic.  Even if you don't have a paid search campaign going, you need to keep an eye on this.  Here's the search engine driven traffic graph for my personal website (www.safdar.net) from Jan. 2009 to Jan 2010:

 

If you look carefully, you'll see a long term trend of decline over the last 12 months with that spike earlier in the year.  This clearly means I need to pay attention to my search engine rank if I want to continue bringing traffic into the site, but even this graph is deceiving.  What I really need to see is the report of which keywords are increasing in traffic, and which are dropping.  That's easily enough done with the "Compare to past" function in Google analytics.   Here's the data for my personal blog:

An article I wrote last year about the season finale of Battlestar Galactica was the biggest driver of traffic, but has since plummeted. (Oh, the nerd in me is revealed!)  This trend on my blog follows the larger, worldwide trend of declining interest in the series now that it's concluded, as I've shown here with Google Insights for Search:

Other phrases are rising.  If I wanted to continually drive traffic to this site, I should examine the phrases that are increasing in traffic, and develop more articles similar to them.  See how that works?  When the results of an analytic measurement can be acted upon, it is actionable.

Step 2: Look deeper at your referral traffic

Your referral traffic, once you remove search engines from it, is a sign of your link-popularity on the web.  Let's look at an example of my referral traffic from Google Analytics with a focus on source and landing page:

When you look at your referral traffic, you can learn what's actually working at sending you traffic.  This shows that some of the images I post get indexed by Google, and that they drive traffic to my website.  (Remember that next time you're writing something, a compelling photo can draw in visitors!)  The knowledge that this drives traffic is what drives me to include posts with my blog posts.  (That's actionable!)

Another thing to look for in this report would be links from other sites, like my first link from www.pokerstove.com.  That link is an enormous driver of traffic for me.  I worked with that site owner to place that link on purpose, but if it happens that I got that link just because he liked an article I wrote, I would want to plan to write more articles of that style, to give him more meaningful things to link to.

Step 3: Rinse, Lather, Repeat (the actionable items)

You don't need to spend a lot of time looking at your analytics every week, just look at what's successfully driving traffic to your site, learn from it, and incorporate that knowledge into your work every week.  Simple actionable things like attaching photos to all your blog posts, identifying keywords that are working, and cultivating relationships with the sites that send you traffic are simple things you can do in a few minutes every week, and will help your long term traffic trends.

Actionable Analytics: Keyword research that brings more traffic to your website

How many times have you had someone give you a "keyword report" from your analytics product and not known what to do with it?  Top keyword lists fall into the realm of analytics data whose purpose is not immediately obvious.  Also in this realm are visitor "traffic reports" and "top content" reports.  Unless you have an active program going to generate traffic from your keywords, then such a report is useless to you.

For example, I once had a client say to me:

Client: "We had xx thousands of visitors come to our website last month.  Is that good?" 

Me: "Do you have any kind of organic marketing program that can influence that number?"

Client: "No."

Me: "Do you have a paid media campaign that gets scaled up or down based upon your traffic?"

Client: "No."

Me: "Do you have any commitments to funders or partners about how much traffic will visit any part of your website?"

Client: "No."

Me: "Then it doesn't matter, you can ignore it for now."

Data like visits, visitors, top content, and top keywords only matter as a barometer for measuring the ongoing work you are doing that may effect them, or if you have some kind of commitment to maintain a particular performance level.  If you have no intention of doing anything to generate more traffic, it's mostly a wasted act to obsess over your traffic reports to learn anything.

Today I'm going to show you where to find the top organic keywords report, which shows what keywords are delivering visitors to your site from search engines.  Then I'll show you how to research these keywords to find similar ones you should be incorporating into your content so that you can get even more traffic.

Of course, every reporting period you can look at the keyword report to measure your progress, which is what makes it an actionable analytic.

Step 1: Find the top keywords/search phrases bringing traffic to your site

This is trivially easy.  On Google Analytics just go to Traffic Sources->Keywords and make sure you click "non-paid".  You should have a report that looks something like this if you select "Landing Page" as the second column item:

If you're using Omniture SiteCatalyst, go to "Traffic Sources->Search Engines - Natural".

Find one or more keywords or phrases that is actually appropriate to you, such as "fundraising metrics" is for me.  If you're constantly writing and publishing, that is an organic search engine optimization program.  How do we find more words similar to this that people search for?

Step 2: Use the Google Wonder Wheel to find similar keywords and phrases

There are indeed a lot of ways to do keyword research.  There are paid tools, and free tools, but one of the first you should look at it the Google Wonder Wheel.  Go take your phrase, like "fundraising metrics" and run it through Google.   Click the "Show Options" link above your search results and you'll see a line of options down the left.  Look all the way down to where it says "Related Searches", and click it.  Your results will look something like this:

From this I picked up the following related search phrases:

1. donation metrics
2. fundraising measurements
3. charity metrics
4. metrics nonprofit fundraising

Copy these down in a list with the original "fundraising metrics" and set it aside.  Now click the Wonder Wheel option right below "Related Searches" and you'll see another way of viewing related searches visually.  (I love the graphical presentation of the Wonder Wheel)  It will look something like this:

The results won't be different, but you can easily expand out to explore related terms.  Don't go too crazy, remember your core term.  After looking at a few of these, I concluded that I should also add "fundraising analysis" to my list.  Let's review the list of phrases:

1. fundraising metrics (the phrase I'm currently using)
2. donation metrics
3. fundraising measurements
   
4. charity metrics
   
5. metrics nonprofit fundraising
6. fundraising analysis

Now we're ready to research them.

Step 3: Evaluate keywords with Google Ad Keywords

The best tool for this is the "Keyword" tool that is built into Google AdWords.  Go get a free AdWords account and find it under "Tools->Keyword Tool".  Copy and paste your list of current and potential phrases into it, and the Keyword tool will tell you incredibly useful things about those phrases, in particular, what is the monthly volume of searches.  Here are the results I got:

This is a great big "DOH!" moment for me.  I have used the phrase "fundraising metrics" in my latest ebook about nonprofit fundraising web analytics without checking to see if that's a highly searched phrase in Google.  Now that I look, I should have used the phrase "fundraising analytics".  That has 10 times more searches than "fundraising metrics" does.

Lesson learned.

Be careful about letting the AdWords tool suggest additional phrases to you.  It casts a very wide net, and you'll find that it suggests phrases that are not appropriate to your content, which can be a waste of time if you're buying ads.  (If you're developing the organic content for it, then it's a fine idea, though it's suggestions can include extremely competitive and generic keywords.)

Once you've identified some phrases that look like they have potential, you can start to incorporate them into your writing, which is the first step of any organic search engine campaign.  Also if you're going to start doing any AdWords purchases, whether its through a Google Nonprofit Grant or through actual budgeted ad dollars, you already know what phrases you want to be advertising on.

Actionable Analytics: 3 ways nonprofits can target a competitor's web visitors with Google Ad Planner

The recent launch of Google for Advertisers is really just a wrapper a number of useful Google tools you already knew about.  But it gave me a good excuse to spend a little time with Google Ad Planner, a good basic tool for figuring out some demographics about who is visiting a website, or how to find the surfing habits of particular demographics.

(To find contrarian views on why you might not want to use Google Ad Planner, go to the end of this document for my "On the other hand" take on the issue.)

There are some really useful things you can do with Google Ad Planner.  I've identified three here, but there are a ton more.

Compare the demographics of the audience of two competing websites

This is really easy.  Go to Google Ad Planner and type in a URL for an organization, like Doctors Without Borders.  I've done this for you.  Notice how it skews for heavily female in its audience, as well as with higher levels of education.  Now compare it to UNICEF.org by following this link.  (Disclosure: UNICEF USA is a client of my firm, Virilion)  Most of the demographics look very similar, until you notice the age breakouts.  UNICEF.org has a significant amount of traffic in the 0-17 year old demographic, whereas Doctors Without Borders has none.

While this may be simply be a problem of audience measurement with Google Ad Planner, it could also indicate that Doctors Without Borders is not developing the next generation of its donor base.  UNICEF has historically done very well in this area, with the Halloween Trick or Treat box and the new J-8 (Junior G-8) summit program targeting younger potential donors.

Remedying this hole should be a priority for Doctors Without Borders.

Find out what other websites a competitor's visitors spend time at

Another great piece of information is the "Sites Also Visited" section of an audience report.  Again, for UNICEF.org, it looks like this:

Notice the correlation of visitors to hrw.org (Human Rights Watch) is 130.0x.  That means that visitors to unicef.org are one hundred and thirty times more likely to visit hrw.org than non-unicef.org visitors.  If you're a competitor of UNICEF.org, you should be approaching Human Rights Watch for partnerships.  From something as simple as featuring non-competitive content in each other's email blasts, to joint projects.  Go where your competitor's customers are, and try and recruit them.

Find out where a specific audience hangs out

Let's say Doctors Without Borders wanted to start targeting 0-17 year olds in the United States by approaching relevant websites for non-monetary partnerships to promote their new "Socially Active High Schoolers" material.  (I'm making that program up, DWB has no such program)  You configure Google Ad Planner to find 0-17 year olds in the United States. 

Then, use the Google Ad Planner filter to select websites that fall into the category "Lifestyle -> Activism and Social Issues". 

What you've got below (I've only shown the top few) is a target list of websites.  You *could* buy advertising on them, but much more affordable (and non-profit-friendly) technique is to approach them to become corporate partners of Doctors Without Borders.  With such a partnership could come many benefits, including access to excess ad inventory or other in-kind donations.  Of course if they really love Doctors Without Borders, they might also run a donation match for their registered users.

And that's just three things...

I've just covered three things, but you can do a lot more with the free audience research of Google Ad Planner.  Don't miss out on an incredibly useful free tool.

On the other hand, why wouldn't I use Google Ad Planner?

Like many Google tools, it's the startup trying to steal market share from more entrenched players.  For anyone who's read The Innovator's Dilemna and the Innovator's Solution, this is a very dangerous competitor.  Most everyone who sells commercial analytics tools has a bigger list of features than the free Google Analytics, but nobody who makes a product in that space enjoys watching the Google Analytics team steal market share as they slowly develop their awesome and free tool from the bottom up.

Google is attacking the problem of advertising and audience research from the bottom.  So it's not going to compete head to head with more advanced services that have been around longer like comScore, but then not everybody needs comScore.  Others have weighed in on this, including famed digital media pundit and player John Batelle.  Wendy Hofstetter also did a piece explaining why they don't compete.

Ultimately, if you're going to drop a serious amount of money in advertising with the sites you find, you're going to migrate to comScore or Nielsen.  If you're not, Google Ad Planner is probably going to be fine, and get better every quarter.

Actionable Analytics: Where is your search engine traffic coming from, and where is it going on your site?

We've all seen this keyword report that shows you the top search engine keywords that drove traffic to your website are.  It's found in Google Analytics under Traffic Sources->Keywords, and once you click "non-paid", it shows you all your organic keywords that are generating traffic from search engines to your website.

This report contains one of the great fallacies of modern web analytics, which is that it defaults to displaying this information ranked on this page by things like visits, pages per visit, and time on site. 

Nobody puts up a website to maximize "time on site" or "visits", they want results!  So click on that tab labeled "Goal Set 1" where you've configured Google Analytics to measure things like sales, e-mail signups, and leads, and then resort your top keywords by "Goal Conversion Rate":

Now this is useful.  Which keywords are most "profitable" for you, yielding the results that justify maintaining your website? 

This is useful information if you're buying keywords because that would tell you how much to pay for them, and whether or not you should buy more or less of them.  But we're talking about organic keywords, now, and organic keywords don't exist on their own.  Every organic keyword exists in tandem with a landing page on your site.  After all, if you like (or don't like) your organic search results, and want to either emulate or improve them, how would you know where to start?  Which of your pages are performing well that you can scrutinize?

The solution comes in a feature rolled out in the latest release of Google Analytics.  Select "Landing Page" from the green-highlighted pulldown menu:

The resulting chart is almost perfect, except that there are a lot of single visits with perfect conversion metrics.  That's too small a sample to draw conclusions from. 

Use the "Advanced Filter" at the bottom to drop out all keyword-landing page combinations that are too small.  While you should use a higher threshhold, I just dropped out every combo with less than 5 visits.  Using the filter is easy, just click "Advanced Filter", and set Visits > 5.

 

.....

The final report is perfect. 

You can use it to look at which combinations of landing pages and keywords are working well, compare those to pages that aren't working well, and either re-edit the copy on the landing pages, change the ask on those pages to raise it's yield, or use the techniques that seem to be working on those pages to create new landing page-keyword combinations.

If you're interested in spending more time on increasing your search engine traffic, and you should, since it's often your best performing traffic, I'll be trying out and reviewing a new product in 2010, Enquisite, to optimize organic search on a number of clients.  (And yes, the people from Enquisite are giving me free access to their tool to try it out.)

The most important Google Analytics report for web/social media managers

The year is ending.  Even most of my marketing is on hold until the beginning of the year.  But if you're responsible for your organization's online presence, including the website, email list, facebook, etc, then you are either:

• agonizing over how to allocate your proposed budget;
• going through the final steps of budget approval; or
• wondering if the budget you proposed has enough money in all the right places.

Regardless of where you are in this process, there is one single Google Analytics report that you must look at right now to tell you how your online efforts are performing.  This report is available in Google Analytics under "Traffic Sources->All Traffic Sources" and will work for you as long as you've got some Goals set up.

Set three variables on this page (I've circled them in red):

• Show: Source/medium;
• View the tab with an active goal set (I've got two sets, so I'm focusing on the second one here); and
• Have it show the conversion rate for one of your key website goals.  I've selected the downloading of my fundraising ebook as the key goal.

Also you'll want to give yourself a good long time frame with the box at the top right.

This report will now show you both the effectiveness of each channel to converting traffic into downloads of my ebook, but also how much that channel is contributing to my overall total.

For example, look at the row labeled "facebook / cpc".  That's for the Facebook cost-per-click ads I run advertising my fundraising ebook.  Those are setup to spend no more than US$5 per day.  They convert at a rate of almost 8% per visit, meaning that for every 100 visits generated, I get 8 ebook downloads.  This campaign generates a whopping 20% of my total conversions, suggesting that I might be able to up my ad budget.

Alternatively, visits coming in from twitter have a very low conversion rate: 0.66%, but because there are so many of them, they actually generate 5% of my total ebook downloads.  Truth be told, I simply post links on twitter when I write something, and occasionally when I find an interesting link.  I'm certainly not spending 5% of my time on it, so it's a very profitable channel, in that respect.

If I was killing myself, posting 30 things to Twitter everyday and still only getting 5% of my total return, I would call it time to scale back or find a more efficient way to do that work.

Feedburner is another example of this.  I don't spend any time actually doing anything special for my RSS feed subscribers, besides writing the content, and yet that is also 5% of my ebook downloads.

This report should tell you how the work you're doing is producing results, and if those results are commensurate with the efforts you're putting forth.

Deploying Omniture SiteCatalyst on an existing website with lots of pages

Having already gotten quite a bit of experience with Google Analytics under my belt (and certified), I've turned to Omniture SiteCatalyst.  The product is vaguely similar to the soon-to-be-decommissioned Hitbox product that my firm has deployed for clients for years, but sort of better in most every way you need it to be.  (More robust, more flexible, etc)  I went thru the three day deployment course, learning the ins and outs of deploying the product and how to configure data collection to get the best possible reports out of it. 

After the course was over, I found one real glaring omission I couldn't believe I had not caught during the course.  It took me a few days to think of an efficient method around it, but it's important enough that I wanted to share it with others, in the spirit of the Internet's open learning environment.

Setup of the Problem Deploying SiteCatalyst

Deployment of SiteCatalyst, at its simplest, is the inclusion of two sets of Javascript tags on every page of your website.  There's one include that references a JS function on your site, affectionately known as the "site code".  The reference looks like this:

<script language="JavaScript" type="text/javascript" src="/js/sitecode.js"></script>

This is no problem, as it's the very same text on every page.  You can include it in a page template for your content management system and it will be included on every page. 

However the second piece of JavaScript on every page is more complicated.  It's longer, and generally looks something like this:

<script language="JavaScript" type="text/javascript">
s.pageName=document.title;  //set the page title

s.server=""

s.channel=""
s.pageType=""
s.prop1=""
s.prop2=""
s.prop3=""
s.prop4=""
s.prop5=""
/* Conversion Variables */
s.campaign=""
s.state=""
s.zip=""
s.events=""
s.products=""
s.purchaseID=""
s.eVar1=""
s.eVar2=""
s.eVar3=""
s.eVar4=""

s.eVar5=""

(At this point you call the function that records the data and sends it back to the Omniture servers, based upon the variables set above)

So, why can't you just stick this into your page template also, and have it included on every page?  Well the variables set on this page actually are page-specific.  Unless you have a good page title set in the <title> tag, you've got to set the pagename.  And if this is a landing page for a specific campaign, you've got to set the "campaign" variable, and various other variables.

Basically, this blog of code is unique for many of your pages.  You can't just shove it into a page template because on a small set of your pages, you're going to want to set some unique traffic and conversion variables.  How was I going to hand edit this code into every page of a thousand+ page site whose content management system already performed slower than I wanted it to?

In fact, how does anyone hand edit this into an existing website with thousands of pages?

When I called Omniture Support, they had two unsatisfying answers.  (This by the way is unusual for them. I've called Support several times and been really pleased with the response and enthusiasm)  The first answer was "Engage Omniture Professional Services".  

The second answer was "hand edit this code into every page of your site" is not an answer that works for any site with more than 100 pages. 

The answer, I discovered, lay in the understanding that for most of my pages, this code is exactly the same.  If I could just figure out a way to include the code on every page in such a way that I could override those defaults for a few select pages, I could save myself the trouble of hand editing all my content, and only edit those pages that have unique variables.

The Solution To Avoid Hand Editing Analytics Code Into All Your Pages

While I'm sure that there are more elegant ways of handling this, one answer lay in the ability of JavaScript to detect the presence of a function, and call it only if it exists.  In short, I realized that I could:

1. Set the variables with default universal values;
2. Check to see if there were any local modifications to those variables on this single page, and if so, set them; and then
3. Call the function to report data to SiteCatalyst as normal.

Once I'd had this epiphany, I realized this code would be able to be universally placed on the templates of my content, letting me avoid the "hand edit every page" problem.  I would then have to hand edit only pages that required specific values to override the defaults.  The result is a block of code that looks like this.  The new addition is in color:

<script language="JavaScript" type="text/javascript">
s.pageName=document.title;  //set the page title

s.server=""

s.channel=""
s.pageType=""
s.prop1=""
s.prop2=""
s.prop3=""
s.prop4=""
s.prop5=""
/* Conversion Variables */
s.campaign=""
s.state=""
s.zip=""
s.events=""
s.products=""
s.purchaseID=""
s.eVar1=""
s.eVar2=""
s.eVar3=""
s.eVar4=""

s.eVar5=""

/* If someone has declared a local function to do some page-specific

  * SiteCatalyst tasks on this page, call it! If they haven't, this is skipped. */

if (typeof window.doLocalSC == 'function') { 

    doLocalSC();

}

(At this point you call the function that records the data and sends it back to the Omniture servers, based upon the variables set above)

Basically I assume that any page that needs special variables will have a JavaScript function "doLocalSC()" defined.  If it doesn't exist, I can continue on and call the SC reporting function.  If it does, I can call it and override the default variables, after which I call the SC reporting function.  Here's an example of one of those local setting functions:

<script language="JavaScript" type="text/javascript">

function doLocalSC() {

s.prop1 = "ebook";

s.campaign = "blahblahblah";

s.pageName=document.getElementsByTagName("h3")[0].innerHTML;

}

</script>

This would be a landing page for my ebook for a marketing campaign called "blahblahblah".  I've set the campaign variable, the traffic property, and also set the pagename a little differently.

This really can reduce the amount of work you need to do to deploy SiteCatalyst, as the main block of code can go into a page template (one edit) and get included on all your content.  Then you only need to concern yourself with hand-editing only those pages with truly unique settings to add "doLocalSC()" functions.

I don't have any idea how Omniture's Professional Services folks would suggest tackling this problem, but that's the solution that seems to greatly reduce the work for me.  I'd be curious to hear other solutions from other fellow analytics nerds familiar with the product.

Slides from the PRNewsOnline "How-To" conference: Online Crisis Communications

My talk at last week's PRNewsOnline "How To" conference at the National Press Club went really well, unless you were the waiter that had to clean up the glass I broke on stage when I knocked it off the podium with my elbow.  The speaker response sheets gave me some of the highest marks of anyone that day, so it had to be entertaining to somebody.

The instruction I was given was to teach people in 20 minutes how to communicate online in a crisis.  I covered some of that, but more than that, I showed the many ways you could turn small disputes into full blown crises.  I assumed that my audience figured out that these were mistakes of others you should learn from, not instructions on how to do it yourself.

The most important thing to learn about communications during a crisis online (or off, for that matter) is something I learned from Ben Popken from Consumerist.com.  He said, during the Mad Magazine-Circuit City kerfuffle that the right response is usually:

1. Admit you did a bad thing;
2. Stop doing the bad thing; and
3. Make a material gesture of apology.

Doing all three is harder than it looks, which is why I find crisis communications so often about changing a company's business practices, rather than about simply explaining those business practices more carefully.

Embedded below are the slides from the talk.  Enjoy!

How nonprofits should measure online advertising

Here are three key lessons every nonprofit should learn and when diving into online advertising. (Classic advertising image from flickr user "pinkponk")

Recently one of my clients purchased a small ad run from a major national newspaper in the Washington DC metro area.  It was roughly $2,000, and they threw in a little creative.  The decision to run the ads and the delivery of all inventory was complete before they let me know. 

This leads us to the first lesson:

Lesson #1: Paid online promotions are experiments.  Start small, with an amount that you feel comfortable testing with.  Make sure it's small enough that if it completely falls flat, you won't beat yourself up, or feel like jumping out the window. 

As we sat down to discuss their latest metrics, my contact turned to me and said "How do we know how it performed?"  One of the reasons I love this particular client is because even though we didn't talk about this particular ad run ahead of time, she used a dedicated campaign URL from Google URL Builder for all the incoming traffic for this campaign.

Lesson #2: If you expect to ever evaluate a promotional campaign of any sort, you need a way of uniquely identifying the traffic.  You often won't be able to segregate it after the fact.  If you use Google Analytics, making a unique URL for your ad campaign can be done with the Campaign URL builder, shown below.

The URL builder is really, really easy to use.  Simply put the URL of the landing page in the box labeled Step #1, fill in the boxes in Step #2 that correspond to your campaign's specifics, and then click the button labeled "Generate URL".  The tool will produce the URL shown in the area highlighted in green.  Copy and paste it into an email to your ad venue and they'll use it.

It's a long and ugly URL, but nobody will ever actually type it, they'll just just be sent to it when they click on your ad.  What's more, it will magically populate your Google Analytics data with all this information.  In fact you can generate as many URLs as you want, using slightly different ones for different banner ads or different campaigns altogether.

This client actually uses one set of URLs for their email list, so they can track the performance of people who click through from the email newsletter.  (It's easy to see why I love them!)

So this ad campaign ran, and after it was over we say down and looked at the performance of the traffic.  We went into Google Analytics and looked at the "Campaigns" section under "Traffic Sources", underlined in red.

When you go there, you get a report like this.  I've attached below a screenshot of a report from my own website (TruthyPR.com) measuring the effectiveness of some different campaign techniques.  Look at the ones I've underlined in red.

One of the big goals of my website is to get people to download my ebook. 

Campaign #10 on the list, simply labeled "Fundraising ebook" is actually a set of Facebook ads I'm paying $5 per day for.  During the time period I've identified it's sent 51 visits to my website, 7.8% of which have downloaded my ebook.  Now look at the Site Average for ebook conversions: 1.48%  That Facebook ad campaign is doing a lot better than average.

However compare it to #6, "How you can help survivors of domestic violence".  That's an email newsletter article I wrote trying to raise money for a domestic violence shelter here in San Francisco that's my personal charity for this holiday season.  Not only did I draw people in to hear my pitch about that, but 2 of those 20 visits also downloaded my ebook in that newsletter last week!

Here's another example out of the Traffic Sources->Search Engines report. 

My cpc ads at Facebook are listed, as are my results from organic google search.  The people that come to my site who happen to find me through google download my fundraising ebook 1.35% of the time.  The people that come to my site through my Facebook cost-per-click ads do it 7.84% of the time.  As the numbers grow I'll have to make a real conscious decision, should I put more money into working on my organic search position, or keep running ads?  These numbers will guide me.

And that brings us to final lesson...

Lesson #3: When evaluating your campaigns, always compare them to other efforts you're undertaking.  Never be fooled into thinking that 7.8% is "good" or "bad".  Focus on who it performs with regards to your other efforts, so you know if you're overspending in those channels.

In my client's case, they found that they got merely a handful of email subscribers from their ad run, and that the cost of the ad run didn't make it economical.  (Really, do you need to pay > $50 per email subscriber?  I doubt it)   More importantly, they found that their Google Grant search ad campaign, which we tweak every month, brings in a higher rate of subscribers and yet the only cost of that is some time to keep it tuned; there's no advertising outlay.

I encouraged them to go back to this media venue and ask them to consider taking on a "cost per action" ad run, that would allow my client only to pay a fixed fee per completed action.  Or, to experiment in another area altogether.  Please indulge me as I suggest one final lesson:

Always experiment.  If you learn something from a poor performing campaign, and do something different to test when you try again, it's worth it if you're using small budgets.  The only crime is in doing the same thing repeatedly and expecting different results.

There are a lot of media companies trying to sell a lot of inventory right now, and many of them are cutting their prices just to get nonprofits in the door as clients and clear some of their unsold inventory.  You're likely to hear from them soon.  You now have the tools to evaluate their results and decide if they're worth the money.

How are my efforts on facebook contributing to my organization's overall goals?

Photo from flickr user clappstar

Introduction

The amount of pressure on organizations to "get on Facebook" is tremendous.  It's driven partly by consultants who use social media expertise as a "guilt marketing tool".  You know what I mean, "You don't understand Facebook, I do, please pay attention to me long enough for me to convince you to hire me!"  I'm guilty of this.

But it's also partly due to the fact that facebook has over 300 million registered users, over 100 million of which use the site monthly.

Those are great numbers, but they aren't "pour money down a hole" numbers.  You need to answer a few simple questions:

1. How effective is the work I do on facebook in producing bottom line results?
2. How should I change my budgeting and resources for facebook to make my results better?
3. Should I take resources away from facebook and devote them to something else?

You can measure the impact your efforts on facebook are having on your bottom line today, and get a good sense of exactly how much time you should be investing in maintaining your outpost there. 

Assumptions About Your Facebook Fan Page

Let's start with what we assume your facebook presence looks like.  I'm guessing it's some sort of fan page on which you post items on your wall, like this one from the Trust for Public Land.  These items lead back to your organization's website through a link.  Maybe, also, they lead to other places as well.  After all, you're not so self-centered as to only share items with your fans that you write.

 

The whole point of your strategy here is to insert your message in front of your fans, mixed in amongst the messages from their friends, so they'll pay attention to it.  In fact, what you really want is for people to click this link I've highlighted in green.

 

If you were to click on this link to the Trust for Public Land website, you'd be taken to this landing page with a little donate link and a description of the project you'd be supporting by clicking the donate link I've highlighted in green.  (For all you optimization folks, yes, that link is too small and doesn't stand out.  TPL is not a client of mine, they just do good conservation work and make for a good example.)

Measuring the impact of your Facebook fan page

While PR measurement experts like KD Paine (one of my heroes) can measure things like "message retention" of your social networking efforts, the easy way to measure the effectiveness of this work on facebook is to track how it hits your bottom line.  When your efforts results in an organization-wide goal (money, volunteers, sponsors) that is easy to understand even by people who don't understand the Internet, there's no debate about the benefit or budgeting of Facebook.

You can create a few metrics from a couple of basic data points:

1. the number of fans you have on a given day you make a fan page status post (gathered from your Fan Page Insights);
2. the number of fans that click on the link to your landing page; and
3. the number of conversions these visits from Facebook generate.

From these data points, you can then build and track the following actionable metrics:

1. clickthru rate of your facebook posts;
2. conversion rate of facebook traffic; and
3. conversion rate of all other "organic" traffic.

Once you have these metrics, you should be able to examine them and look to connect improvements or degradations in the results to specific actions. Do all my status updates without a photo perform worse than those with a photo of the land being conserved? Then we should always use a photo!

Objections to this measurement strategy

Your first reaction to seeing this might be "Hey, wait, I've got lots of people clicking 'Like' on my status updates, and 'Share' and playing my videos.  Doesn't that count for something?"

Well, to be concise about it, not really.

Yes, facebook is different. Yes you have to hold conversations with your fans, you can't just broadcast to them. I'm not advocating that you put a donate link up every day with no context. Your facebook presence needs to be more interactive than that, with you contributing to a larger conversation by highlighting other areas that are important to your audience, and bringing other things besides calls-to-action into the relationship.

And most importantly, listening to what people are saying (on your Facebook fan page and off) and then incorporating that into your work.

But in the absence of results, it doesn't matter how many people click 'Like' on your status update.  It doesn't matter how many people share your video.  It doesn't even matter how many fans you have. 

You don't have results.  Of course, once you have results, you have the luxury of measuring other things more subtle, like non-transactional interactions.  But since you don't know if you have results yet because you're reading this ebook to learn how to measure them, you don't have that luxury.

Start with measuring well-understood results like money and volunteers, and optimizing your status updates and landing pages to increase those results.  I'll talk about measuring non-transactional interactions in future writing.

Pulling it all together

For this example we're assuming the top goal of your website is fundraising, but you could apply this equally well to advocacy efforts where you want your visitors to write letters to Congress, e-commerce efforts where you want to sell products, etc. 

I'll explain more about how to collect this data in a future post, but suffice to say, you'll need to pull it from two sources: your analytics product on your website and Facebook's Fan Page Insights.

Here's an example spreadsheet of the data you should create around your status messages:

Once you've done that, you can aggregate it up for weekly summaries and compare the results to the rest of your site traffic.  I suggest that you focus on weekly batches of data.  I find that weekly meetings with my clients creates a good rhythm that allows us enough data to see the results of our experiments, and gives us enough time to put in changes for new experiments. Here's a spreadsheet where I've done that with more imaginary data I made up just as an example.

Where to go from here?

Once you've got a trove of data from at least 50 updates and 4 weeks, you'll need to start optimizing your facebook updates.  Some common areas for optimization are:

1. COPY: Are there common elements in the title or copy that cause more people to click?  
2. URGENCY: How much higher is my clickrate when I incorporate a sense of urgency into the call to action?
3. DAYPARTING: Does the day of the week or time of day matter?  (You can look at Facebook Fan Insights to get a sense of the timezone your fans are in, and even use their targeting tools to show status updates only to certain timezone's of people.)

You should always be either setting up an experiment or collecting data from one to figure out whether your hypotheses are working.  Good luck, and may the results set you free.