Friday, October 14, 2011
Google Analytics Blog: Webmaster Tools in Google Analytics for everyone
Google Analytics Blog: Webmaster Tools in Google Analytics for everyone: Back in June, we announced a pilot program to allow users to surface Google Search data in Google Analytics by linking their Webmaster Tool...
Wednesday, July 27, 2011
Customizing Omniture
Recently, I've been using Omniture er um I mean Adobe Analytics to extract some rather complex metrics. A few things struck me that I believe I should share. Because I've tried not to blog about too many technical or implementation issues, I thought I would explain how Omniture has been providing more and more new features which are reducing the need to "re-implement" new code in the form of new (jargon alert) e-vars.
New Implementation:
This is really targeted to those of you who (like me) remember the horror stories (or worse yet lived the nighmare) of a painful omniture implementation. Omniture now claims that they are leveraging the experience from deploying SiteCatalyst over 3000 times and have developed best practices which combine business, industry, and technical needs in a process they are calling their Fusion Methodology.
If you are wondering how much easier this is since the "good ol days," consider that Omniture is still suggesting that implementation will take between 4 and 6 weeks- working with a technical contact and includes weekly status calls. It's still not as easy as GA for sure. They have also moved from what they call "G" code to "H" code and offer migration guidelines.
Server Side versus Client Side scriptng also seems to be an important "new" option. It used to be that all the code was on the client side and that was the only option. That is based on hearsay on my part, but my research seems to back that up as well. I never personally implemented under this methodology. The "variables" can now also be contained in a library for reference by the javascript.
This is in no way designed to be an exhaustive list of all implementation methodologies, but suffice it to say that if you have been constantly ruling out Omniture because you don't have the technical resources or the time to implement this robust solution, you may want to look again.
Modifications:
Depending on how your particular instance of Omniture was implemented (and likely WHEN), you may find yourself lacking the flexibility you might "want" in the reporting suite. Traditionally, the problem I've run against is that a website is a dynamic entity which often doesn't resemble itself from one year to the next. Check out the archive.org project and you'll see what I mean. Also- all too often, you "don't know what you don't know" when you first implement any analytics application and that can be a real problem.
Omniture seems to have been adding functionality to help with these issues. Custom dashboards are easier that ever to build and their Excel tools are greatly improved from the past as well. The "Excel Client" has been replaced with the new "Report Builder" which opens up a whole new layer of flexibility. Originally, I was not a fan of "having" to use excel with an enterprise application like Site Catalyst, but it turns out that there is tremendous benefit.
For one thing, most business users are already trained on how to use excel. It's a native environment for them. Also- regardless of your level of experience with Site Catalyst, custom dimensions and metrics can be confusing and you're likely to run into limitations. Report Builder removes many of those limitations. It also provides a means of refreshing the data automatically and can even send reports on a schedule as well.
I recently ran into a situation where the only way I could determine to reasonably get the data I wanted was by using Report Builder. Unfortunately, I needed to request access to Report Builder from the admin of this client- which can sometimes be a problem. It's also kinda hidden- I couldn't find it even when I knew what I was looking for. There doesn't appear to be a way to just download the excel add-in except from the post log in menu and even then, it won't do you much good if you don't have the proper permissions to use it.
Perhaps the biggest benefit to Report Builder is that, based on my unscientific observations, I'd say it would likely dramatically reduce the need to re implement code or even new variables into the library. That removes a ton of hassle and, of course, the need to wait for new data to build- since you can't go back in time with new variables.
Remember- approach your analytics in the right order:
Re implement only if you're sure it's necessary...
New Implementation:
This is really targeted to those of you who (like me) remember the horror stories (or worse yet lived the nighmare) of a painful omniture implementation. Omniture now claims that they are leveraging the experience from deploying SiteCatalyst over 3000 times and have developed best practices which combine business, industry, and technical needs in a process they are calling their Fusion Methodology.
If you are wondering how much easier this is since the "good ol days," consider that Omniture is still suggesting that implementation will take between 4 and 6 weeks- working with a technical contact and includes weekly status calls. It's still not as easy as GA for sure. They have also moved from what they call "G" code to "H" code and offer migration guidelines.
Server Side versus Client Side scriptng also seems to be an important "new" option. It used to be that all the code was on the client side and that was the only option. That is based on hearsay on my part, but my research seems to back that up as well. I never personally implemented under this methodology. The "variables" can now also be contained in a library for reference by the javascript.
This is in no way designed to be an exhaustive list of all implementation methodologies, but suffice it to say that if you have been constantly ruling out Omniture because you don't have the technical resources or the time to implement this robust solution, you may want to look again.
Modifications:
Depending on how your particular instance of Omniture was implemented (and likely WHEN), you may find yourself lacking the flexibility you might "want" in the reporting suite. Traditionally, the problem I've run against is that a website is a dynamic entity which often doesn't resemble itself from one year to the next. Check out the archive.org project and you'll see what I mean. Also- all too often, you "don't know what you don't know" when you first implement any analytics application and that can be a real problem.
Omniture seems to have been adding functionality to help with these issues. Custom dashboards are easier that ever to build and their Excel tools are greatly improved from the past as well. The "Excel Client" has been replaced with the new "Report Builder" which opens up a whole new layer of flexibility. Originally, I was not a fan of "having" to use excel with an enterprise application like Site Catalyst, but it turns out that there is tremendous benefit.
For one thing, most business users are already trained on how to use excel. It's a native environment for them. Also- regardless of your level of experience with Site Catalyst, custom dimensions and metrics can be confusing and you're likely to run into limitations. Report Builder removes many of those limitations. It also provides a means of refreshing the data automatically and can even send reports on a schedule as well.
I recently ran into a situation where the only way I could determine to reasonably get the data I wanted was by using Report Builder. Unfortunately, I needed to request access to Report Builder from the admin of this client- which can sometimes be a problem. It's also kinda hidden- I couldn't find it even when I knew what I was looking for. There doesn't appear to be a way to just download the excel add-in except from the post log in menu and even then, it won't do you much good if you don't have the proper permissions to use it.
Perhaps the biggest benefit to Report Builder is that, based on my unscientific observations, I'd say it would likely dramatically reduce the need to re implement code or even new variables into the library. That removes a ton of hassle and, of course, the need to wait for new data to build- since you can't go back in time with new variables.
Remember- approach your analytics in the right order:
- Strategy
- Implementation (Access)
- Reporting
- Ad Hoc Analysis
Re implement only if you're sure it's necessary...
Thursday, March 24, 2011
Connecting Adwords and Google Analytics
I don't usually blog about "technical" issues, so I'll try to keep this as practical as possible. There are a few really important decisions that need to be made when setting up both Adwords and Google Analytics. Some of these decisions can have traumatic long term implications by missing a few key subtleties.
Beginning with the end in mind, suppose you have a few websites. Many companies do. You're likely to want to manage those websites separately, but you may also want to see them "together" in a roll up. So- many Google Analytics Gurus will recommend setting up ONE account with several "profiles" to see the separate websites. That's the easiest way to set things up. With a few tweaks, you can even see traffic crossing domains and referring to one another.
This can have some unintended consequences. You are also likely to want to run separate AdWords campaigns with separate line item budgets. In order to do that, you'll need at LEAST separate campaigns (because you can only control budgets at the campaign level) but ideally you may even need separate ACCOUNTS.
As of a few years ago, you can now "connect" Adwords and Analytics. that's a really nice feature- you can open analytics directly in adwords, and some other nice features present themselves as well- likely even more to come as this becomes more standard. Here's the problem:
You can't connect one Adwords account with a "profile" you can only connect an adwords account with one analytics account. To make matters even worse, some Analytics accounts are set up with an incrementing "UA" number in other words, you can set
up a new "site" under the same account and your UA structure will look something like this: UA-XXXXXX-1 then UA-XXXXXX-2 and so on. Google Analytics mixes it's metaphors on this one- and they consider these to be the same account but separate profiles- even though there are unique profile numbers assigned separately. Perhaps Danny Sullivan will add this to his list of the 25 Things I Hate About Google and this one may be on it next time.
This may seem rather trivial, but it's pretty interesting the number of things that don't behave properly as a result of this. So how do you fix it? Well- that's the worst part- the only fix is to start new accounts- which of course means starting your numbers from scratch...and this is according to two separate Googlers I've spoken with about this issue. It's kinda one of those things where you need to think carefully about whether there are enough reasons to re-implement. Unfortunately, there may be more and more features only available for accounts that are connected.
If your accounts are connected, You'll get this neat option inviting you to import conversions from Google Analytics into adwords under "conversions." Only problem is, I've tested this a bunch, and it's unreliable. I'm not even allowed to try it again or my boss will get mad at me. He's convinced it's never going to work. Ironically, you can't import Adwords conversions into Google Analytics. It's just not an available feature...so you have build conversions in Adwords and goal conversions in GA...urgh.
So- be careful when setting up your new Google Analytics account and reach out to me if you need info on some of the work arounds we've developed- there out of the scope of the blog.
Beginning with the end in mind, suppose you have a few websites. Many companies do. You're likely to want to manage those websites separately, but you may also want to see them "together" in a roll up. So- many Google Analytics Gurus will recommend setting up ONE account with several "profiles" to see the separate websites. That's the easiest way to set things up. With a few tweaks, you can even see traffic crossing domains and referring to one another.
This can have some unintended consequences. You are also likely to want to run separate AdWords campaigns with separate line item budgets. In order to do that, you'll need at LEAST separate campaigns (because you can only control budgets at the campaign level) but ideally you may even need separate ACCOUNTS.
As of a few years ago, you can now "connect" Adwords and Analytics. that's a really nice feature- you can open analytics directly in adwords, and some other nice features present themselves as well- likely even more to come as this becomes more standard. Here's the problem:
You can't connect one Adwords account with a "profile" you can only connect an adwords account with one analytics account. To make matters even worse, some Analytics accounts are set up with an incrementing "UA" number in other words, you can set
up a new "site" under the same account and your UA structure will look something like this: UA-XXXXXX-1 then UA-XXXXXX-2 and so on. Google Analytics mixes it's metaphors on this one- and they consider these to be the same account but separate profiles- even though there are unique profile numbers assigned separately. Perhaps Danny Sullivan will add this to his list of the 25 Things I Hate About Google and this one may be on it next time.This may seem rather trivial, but it's pretty interesting the number of things that don't behave properly as a result of this. So how do you fix it? Well- that's the worst part- the only fix is to start new accounts- which of course means starting your numbers from scratch...and this is according to two separate Googlers I've spoken with about this issue. It's kinda one of those things where you need to think carefully about whether there are enough reasons to re-implement. Unfortunately, there may be more and more features only available for accounts that are connected.
If your accounts are connected, You'll get this neat option inviting you to import conversions from Google Analytics into adwords under "conversions." Only problem is, I've tested this a bunch, and it's unreliable. I'm not even allowed to try it again or my boss will get mad at me. He's convinced it's never going to work. Ironically, you can't import Adwords conversions into Google Analytics. It's just not an available feature...so you have build conversions in Adwords and goal conversions in GA...urgh.
So- be careful when setting up your new Google Analytics account and reach out to me if you need info on some of the work arounds we've developed- there out of the scope of the blog.
Tuesday, March 22, 2011
Thursday, January 6, 2011
2011 the Year of Dividends
I've been reading all sorts of articles and blog posts these days offering tips for the New Year. I thought it would be best to offer a few of my consolidated thoughts from my various explorations and my own experiences.
- ROI- 2011 should be the year that our customers are able to more accurately measure their return on investment. Web analytics is very powerful when it is applied to real life business decisions. I am still surprised and how few marketers still don't know how their marketing efforts affect their bottom line. This is often a significant challenge, but new technology and more sophisticated techniques can bridge that gap. Get there. It's worth the effort. Without being able to measure ROI, you won't be able to understand how to cash in on your marketing efforts.
- Cash in your Analytics Dividends- If you've been "analysing" a lot, the first part of the year may be a good time to cash in on some learnings. This can take several forms, but here is an example. We know of a company that has invested a considerable amount of money in PPC advertising. Using analytics as a guide, they can now expand into new markets or at least new market segments while applying those learnings. The pain and suffering is over and now they can count on a lower cost per conversion.
- Refine your Dashboards- If you are still feeling overwhelmed by all the data you are dealing with, take advantage of the built in "first of the year" recalibration excuse. Have that conversation with your team about simplifying your life by focusing on the KPI's (key performance indicators) that really matter to your bottom line.
Tuesday, November 16, 2010
3 reasons why this is a BAD Link request
I'm not sure why I'm SO surprised that it has been a while since I've received such a blatant and poorly executed attempt to exchange links. Perhaps I thought that most people were wiser by now. For those who still haven't heard, this is NOT how you exchange links:
Dear Sir or Madam
We would sincerely like to exchange links with your website.
Exchanging links will increase search engine rankings and website popularity, therefore, we will both benefit from exchanging links.
If you are interested, please add our link on your site and let me know.
Our linking details are as follows:
Title: Stourbridge Motor Company
URL: http://www.stourbridgemotorcompany.co.uk
Description: Stourbridge Motor Company deliver quality used cars, vans and mpv’s at realistic prices in West Midlands, Birmingham, Stourbridge, Redditch, Halesowen, Dudley and surrounds.
Please let me know as soon as you have posted our site details onto a link page that is cached by Google.
In exchange, please send your link information to me, which we will upload within 24 hours of receipt, and we will also notify you as soon as the link is posted.
We sincerely forward to your positive response.
Kind regards
[from someone I've never heard of]
Three reasons why this is wrong:
1. My site has nothing to do with their site: They're even on a different continent!
2. Trading links for pure SEO value is a huge NO NO. Even if that's what you're really doing, that's not why you should be doing it. You should exchange links because your site and the site you're exchanging a link with are in the same "Cyber neighborhood." Otherwise, Google won't understand the connection and it does neither of you any good.
3. Unsolicited link exchange requests are still SPAM.
Wednesday, May 26, 2010
Google Analytics for SEO
I recently had the opportunity to speak on the topic of analytics at SMX Advanced in London. The topic of our panel was Making Sense Of All The Numbers. Since it was the last session on the second day, I was concerned that perhaps it wouldn't have been as well attended as it was. However, as the description indicated, We attempted to advocate various methods that could be implemented in order to maximize the investment of time and money that you have likely made in web analytics.
While I plan to share more details about my presentation, I first wanted to share a few Ah Ha moments that I had during the course of the conference. The first day of the conference was divided between an SEO and a PPC track and I switched back and forth. The first Ah Ha moment was really contained inside the presentation made by Barak Berkowitz from WolframAlpha during which I could hardly contain myself with interest and awe at this new "Calculation engine" or Knowledge Engine as they describe it. It's very cool- check it out. What was apparent however is that it's coolness is NOT immediately apparent. The main reason for that is that it is not a search engine. they need to do a better job of demonstrating it's benefits to new visitors.
The second Ah Ha moment was really more of a reinforcement of some previously less than obvious realities about Google Analytics. GA is not designed for measuring SEO. It is still a website analytics package that falls short in some important areas and this is one of them. It can certainly be "hacked" into shape- but "out of the box" lots of people complain about this in forums like these, but try to be careful never to offend any Googlers lingering about.
If you need help customizing Google for SEO, consider building a separate profile and segment all your traffic to organic. It's pretty easy to do, but you're likely to need to wait a little while more traffic builds in the new profile. I also followed some interesting threads and have experimented with getting "rankings" into the reports. Pretty neat stuff. I know more about what Google's up to, but I've been cautioned not to share any of this yet or I'll get kicked out of the Beta club.
While I plan to share more details about my presentation, I first wanted to share a few Ah Ha moments that I had during the course of the conference. The first day of the conference was divided between an SEO and a PPC track and I switched back and forth. The first Ah Ha moment was really contained inside the presentation made by Barak Berkowitz from WolframAlpha during which I could hardly contain myself with interest and awe at this new "Calculation engine" or Knowledge Engine as they describe it. It's very cool- check it out. What was apparent however is that it's coolness is NOT immediately apparent. The main reason for that is that it is not a search engine. they need to do a better job of demonstrating it's benefits to new visitors.
The second Ah Ha moment was really more of a reinforcement of some previously less than obvious realities about Google Analytics. GA is not designed for measuring SEO. It is still a website analytics package that falls short in some important areas and this is one of them. It can certainly be "hacked" into shape- but "out of the box" lots of people complain about this in forums like these, but try to be careful never to offend any Googlers lingering about.
If you need help customizing Google for SEO, consider building a separate profile and segment all your traffic to organic. It's pretty easy to do, but you're likely to need to wait a little while more traffic builds in the new profile. I also followed some interesting threads and have experimented with getting "rankings" into the reports. Pretty neat stuff. I know more about what Google's up to, but I've been cautioned not to share any of this yet or I'll get kicked out of the Beta club.
Tuesday, March 23, 2010
What Agencies Want from Marketers
It occurred to me that I've had some valuable experiences in having presented new marketing concepts and analytics packages to both Ad Agencies and Digital Marketing Agencies. It also occurred to me that perhaps others may value some of my insights when considering a quest to gather agency buy in of a new idea.
As mentioned in other previous posts, agencies can't help being very practical. The days of hugely profitable, hardly measurable percentage based revenue models are nearly gone. In years past, it was common for agencies to "easily" earn 15% of an advertiser's entire media budget in fees. But as advertising becomes more digital and more measurable, more and more advertisers began questioning the logic behind this method. It's now much more common for agencies to negotiate a fee based or "retainer" type of an approach. As a result, agencies have been migrating toward greater efficiency and toward ideas that will generate additional revenue.
Digital agencies (especially larger ones) should care more about what works and what doesn't but also tend to favor "win-win" low or no "cost" programs. Agencies certainly don't mind selling their own wares including analytics solutions, ad serving technology, ad delivery "networks" and of course creative services.
In having spoken with hundreds of agencies about interactive marketing, I have observed the following common ground:
1. Agencies are not as creative as you may expect them to be. They seem to want to crave new ideas, but you'd better put your best foot forward because they don't want to have conversations about concepts.
2. Agencies will pick the lowest hanging fruit. If you make their life easier and you make them money- you're in. If it actually works for their customers, that's an added bonus that we help you stick around beyond an initial trial.
3. Agencies will ask for a trial. They rarely "buy" anything to begin with, but they hardly even buy anything before they've had a chance to test it.
4. Agencies won't actually do a fair trial. They will forget, get too busy, not fully implement (especially not according to your desired specs) They'll say they will, but they won't
5. Ad Agencies will ask for the lowest price possible. Then they'll ask again. then they'll refer you to someone else who will ask for an even lower price.
6. Digital Marketing Agencies will require excellent results and their standards will start off very high. Can you blame them? Many sold their clients media void of any real metrics for years.
7. Analytics Agencies (or an agency with a strong analytics focus) will ask for integrated metrics and conversion analytics.
8. Noone wants to be the first to present an idea to a client, but ironically, they don't want anyone else talking to their client either.
9. They won't introduce you to their clients unless they are fully confident that there is no risk at all that you will have any chance of working with them directly. In other words, they won;t introduce you to their clients.
10. Even if you present case studies, agencies will want a case study from an exact clone of their client. If you have a B2B example, they'll want a B2C example. If you're case study is for a Dentist, they'll complain that it's not for a Chiropractor. It's inevitable.
All this being said, establishing strong relationships with Agencies can be a very important goal. However, be careful that you don't burn bridges by jumping in too soon. You may only get one shot to prove your value proposition.
As mentioned in other previous posts, agencies can't help being very practical. The days of hugely profitable, hardly measurable percentage based revenue models are nearly gone. In years past, it was common for agencies to "easily" earn 15% of an advertiser's entire media budget in fees. But as advertising becomes more digital and more measurable, more and more advertisers began questioning the logic behind this method. It's now much more common for agencies to negotiate a fee based or "retainer" type of an approach. As a result, agencies have been migrating toward greater efficiency and toward ideas that will generate additional revenue.
Digital agencies (especially larger ones) should care more about what works and what doesn't but also tend to favor "win-win" low or no "cost" programs. Agencies certainly don't mind selling their own wares including analytics solutions, ad serving technology, ad delivery "networks" and of course creative services.
In having spoken with hundreds of agencies about interactive marketing, I have observed the following common ground:
1. Agencies are not as creative as you may expect them to be. They seem to want to crave new ideas, but you'd better put your best foot forward because they don't want to have conversations about concepts.
2. Agencies will pick the lowest hanging fruit. If you make their life easier and you make them money- you're in. If it actually works for their customers, that's an added bonus that we help you stick around beyond an initial trial.
3. Agencies will ask for a trial. They rarely "buy" anything to begin with, but they hardly even buy anything before they've had a chance to test it.
4. Agencies won't actually do a fair trial. They will forget, get too busy, not fully implement (especially not according to your desired specs) They'll say they will, but they won't
5. Ad Agencies will ask for the lowest price possible. Then they'll ask again. then they'll refer you to someone else who will ask for an even lower price.
6. Digital Marketing Agencies will require excellent results and their standards will start off very high. Can you blame them? Many sold their clients media void of any real metrics for years.
7. Analytics Agencies (or an agency with a strong analytics focus) will ask for integrated metrics and conversion analytics.
8. Noone wants to be the first to present an idea to a client, but ironically, they don't want anyone else talking to their client either.
9. They won't introduce you to their clients unless they are fully confident that there is no risk at all that you will have any chance of working with them directly. In other words, they won;t introduce you to their clients.
10. Even if you present case studies, agencies will want a case study from an exact clone of their client. If you have a B2B example, they'll want a B2C example. If you're case study is for a Dentist, they'll complain that it's not for a Chiropractor. It's inevitable.
All this being said, establishing strong relationships with Agencies can be a very important goal. However, be careful that you don't burn bridges by jumping in too soon. You may only get one shot to prove your value proposition.
Tuesday, February 2, 2010
Have You Tried the "Wonder Wheel"?
Of all the cool features Google offers, the Wonder Wheel is one of the nicest hidden gems. Especially for those who are interested in keyword research. It runs as a flash object inside the Google Results. To find it, click the +show options link on the top left side of the search result page. The "t
ool belt" (Matt Cutt's term) will show on the left hand side and click on Wonder Wheel. Here are sample results. Notice that you can select any related term and click on it to see the related wheel. This is a great keyword research tool. I still haven't figured out a way to export this info, so you may need to open an excel spreadsheet or bust out a pencil.
While you're exploring, check out other tool belt features. For instance, notice that you can search by time frame. this is great for repuation management. You can search for what's been said about you or your company within the past week or even the past 24 hours.
You can also see search results limited to blog posts only etc.
As always, Google is inventing new ways for users to find good information and along the way, there are beneifts to professional search marketers as well.
ool belt" (Matt Cutt's term) will show on the left hand side and click on Wonder Wheel. Here are sample results. Notice that you can select any related term and click on it to see the related wheel. This is a great keyword research tool. I still haven't figured out a way to export this info, so you may need to open an excel spreadsheet or bust out a pencil. While you're exploring, check out other tool belt features. For instance, notice that you can search by time frame. this is great for repuation management. You can search for what's been said about you or your company within the past week or even the past 24 hours.
You can also see search results limited to blog posts only etc.
As always, Google is inventing new ways for users to find good information and along the way, there are beneifts to professional search marketers as well.
Friday, October 23, 2009
Google Analytics newest features
I pulled out my flip video during a short breakout session at eMetrics and want to share a fun and informative presentation by Phil Mui and Avinash Kaushik from Google Analytics. For the past few years, Google Analytics has steading added a surprisingly nice set of additonal features. I'm proud to say that I recognize many of these features from the "good ol days" at ClickTracks. Avinash was the lead analyst at Intuit at the time and I still remember when the announcement came that he would become the Web Analytics evangelist at Google Analytics. With the latest features, Google Analytics is better than most other commercially available products. It's gone from "great for the price" (free) to worth more than free. I think they will continue to damage uptake of commercially available and often very expensive solutions that are left with fewer differentiators.
Take a look:
Part 2:
Part 3:
Take a look:
Part 2:
Part 3:
eMetrics Summit in Washington DC
I just returned from an exciting three days in DC at the Web Analytics Association's eMetrics Summit. Jim Sterne opened with a great keynote. The theme was "Translating Web Intelligence into Business Value." That's a great goal and certainly a major theme of professional Web Analysts. The four main points he stressed was a focus on measuring:
Behavior or what do visitors "DO"
Attitude or what do visitors "FEEL"
Competition or how does your website "COMPARE"
Outcomes or what are the "RESULTS" of your efforts online
He stressed that management typically wants more than just data. They want your opinion.
He also encourages analyst to "Reason backwards" when building funnel analysis. In so doing, you can avoid a common trap in analysing data which is faulty causality reasoning. An example would be if you were to notice that you are selling many more umbrellas on days when it is raining, it doesn't mean that umbrellas cause rain.
He offered a "gift" to attendees which is a list of the 101 things you can do.
Beyond the keynote, there were lots of additional sessions during which I took good notes and built my own knowledge base tremendously. It's likely I'll blog about other topics soon.
Behavior or what do visitors "DO"
Attitude or what do visitors "FEEL"
Competition or how does your website "COMPARE"
Outcomes or what are the "RESULTS" of your efforts online
He stressed that management typically wants more than just data. They want your opinion.
He also encourages analyst to "Reason backwards" when building funnel analysis. In so doing, you can avoid a common trap in analysing data which is faulty causality reasoning. An example would be if you were to notice that you are selling many more umbrellas on days when it is raining, it doesn't mean that umbrellas cause rain.
He offered a "gift" to attendees which is a list of the 101 things you can do.
Beyond the keynote, there were lots of additional sessions during which I took good notes and built my own knowledge base tremendously. It's likely I'll blog about other topics soon.
Wednesday, October 7, 2009
SMX East takeaways
I've been attending Search shows for several years. I'm at T5 using their stable wifi (unlike the service available at the Javit's Center) while I wait on my delayed flight to board. I ran into Eric Enge in the Hall and introduced him to Ryan Kelly with Pear Analytics. Eric described me to Ryan as "coin operated." I laughed...then quickly realized that perhaps I should have been offended. Then just as quickly realized he didn't mean it as an insult. I then reminded eric that I had successfully extracted some coins from him in the past so I decided that it was more likely a high compliment. Eric is a Rock Star SEO and I feel like I've contributed to his analytics prowace.
There were fewer vendors and mostly the same cast of characters, but missing several notables. There were fewer booths for sure which is likely a sign of the times but not likely the industry. Others shared my excitement about the SEO business in particular. More and more marketers are catching on to the benefit. Rand Fishkin from SEOmoz threw out a zinger during a session this morning when he pointed out to a guy who's running a bunch of PPC, "you know google doesn't charge you for those clicks over there under those paid ads." The guy was there to learn about SEO and knew that, so the comment was met with a big laugh and the guy was not offended in the least since he was in on the joke.
It strikes me that these seminars are great for people who already "get" seo/ppc, but not a training course by any means. The panel format is often tricky- some are unprepared, others are long winded. In other cases, it's exactly the right approach. The first panel had me very worried when 5 of the most respected SEO's gave completely conflicting "opinions" about the use of NoFollow tags as it related to passing link juice. The aforementioned Eric Enge presented the only real case study data on the topic as Rand admitted he wasn't done studying the issue yet.
Because I'm commited to simplfying analytics, I'll try to sum up the issue and relate it to Analytics.
NoFollow is a Meta (behind the scenes) coding attribute which basically tells the search engines not to follow a specific link from a website. For a while, SEO practitioners were convinced that if a site has a bunch of links to external pages, the site would bleed "link Juice" which is a credibility rating issued by Google. (oversimplified for sure...)
So- ack in May, as the story goes, Matt Cutts (Google's SEO one man army) explained that the reasons SEO's had been using NoFollow was invalid. So- while this may not be directly related to analytics, Eric used analytics to show an increase in total indexed pages and total traffic right after one of his clients removed their nofollow tags.
My head was spinning as others explained other methods of "link sculpting" which is a concept noone seems to want to associate themselves with...UNLESS it's used for improving user experience instead of gaming the search engines. So the verdict seemed to be, "it depends" which almost always causes audience groans.
As happens, hallway conversations cleared up my thoughts on the idea and I'm now of the mind that nofollow sucks. If you don't want Google to follow your link, don't put the link on your page.
While I don't have time to blog all the feedback from the show, I have idea foder for several more posts over the next few days. But before I go, I found a really cool new analytics package which made it's debute at the show. It was created by an SEO turned tool company called Pear Analytics. Their new product is called Site Juice and they offer a 30 day trial and a tour on their site. They are doing some really cool things with NO CODE install required. May seem like a shameless plug, but I like their ideas and their commitment to simplicity.
WOAH- My flight went from delayed to ON TIME! Gotta go...
There were fewer vendors and mostly the same cast of characters, but missing several notables. There were fewer booths for sure which is likely a sign of the times but not likely the industry. Others shared my excitement about the SEO business in particular. More and more marketers are catching on to the benefit. Rand Fishkin from SEOmoz threw out a zinger during a session this morning when he pointed out to a guy who's running a bunch of PPC, "you know google doesn't charge you for those clicks over there under those paid ads." The guy was there to learn about SEO and knew that, so the comment was met with a big laugh and the guy was not offended in the least since he was in on the joke.
It strikes me that these seminars are great for people who already "get" seo/ppc, but not a training course by any means. The panel format is often tricky- some are unprepared, others are long winded. In other cases, it's exactly the right approach. The first panel had me very worried when 5 of the most respected SEO's gave completely conflicting "opinions" about the use of NoFollow tags as it related to passing link juice. The aforementioned Eric Enge presented the only real case study data on the topic as Rand admitted he wasn't done studying the issue yet.
Because I'm commited to simplfying analytics, I'll try to sum up the issue and relate it to Analytics.
NoFollow is a Meta (behind the scenes) coding attribute which basically tells the search engines not to follow a specific link from a website. For a while, SEO practitioners were convinced that if a site has a bunch of links to external pages, the site would bleed "link Juice" which is a credibility rating issued by Google. (oversimplified for sure...)
So- ack in May, as the story goes, Matt Cutts (Google's SEO one man army) explained that the reasons SEO's had been using NoFollow was invalid. So- while this may not be directly related to analytics, Eric used analytics to show an increase in total indexed pages and total traffic right after one of his clients removed their nofollow tags.
My head was spinning as others explained other methods of "link sculpting" which is a concept noone seems to want to associate themselves with...UNLESS it's used for improving user experience instead of gaming the search engines. So the verdict seemed to be, "it depends" which almost always causes audience groans.
As happens, hallway conversations cleared up my thoughts on the idea and I'm now of the mind that nofollow sucks. If you don't want Google to follow your link, don't put the link on your page.
While I don't have time to blog all the feedback from the show, I have idea foder for several more posts over the next few days. But before I go, I found a really cool new analytics package which made it's debute at the show. It was created by an SEO turned tool company called Pear Analytics. Their new product is called Site Juice and they offer a 30 day trial and a tour on their site. They are doing some really cool things with NO CODE install required. May seem like a shameless plug, but I like their ideas and their commitment to simplicity.
WOAH- My flight went from delayed to ON TIME! Gotta go...
Thursday, July 9, 2009
Negotiation Nation
This video is just too funny! Part of what I find funniest about it is that I have been "jokingly" trying some of these techniques for years. The check out person at Walmart usually finds it funny as soon as they realize I'm joking. For some reason, lots of business people have become accustomed to being able to negotiate and will get it done. In some cases, it can make some sense to leverage into the future or try to work within a budget, but it can be very frustrating when the negotiation seems to be just "for sport."
Tuesday, July 7, 2009
Online Reputation Management
Here's a term that's been thrown around a lot these days. In fact, when I mention to anyone that my company helps with Online Reputation Management, they inevitably quiz me on the topic. What is it? How does it work?
Online Reputation Management is the practice of understanding the buzz about a product, person, or even an offline company. An online reputation is just as often good as it is bad. In fact, true online reputation management is essentially public relations in the digital world of instant communications.
There are lots of similarities to the standard tenets of Public Relations, but many phenomenon have been amplified. "It's been said" and many have learned first hand that one positive comment made from one person to another may impact a sale in a positive way. Offline, that may translate to a 1 to 1 ratio. Online, if the information is "shared" through social media etc., it may create a 4:1 ratio (4 sales to every 1 positive comment) as a result of a "viral spread." The dangerous part is the amplification of the negative feedback loop.
In the offline world, the basic marketing principle is that for every one negative comment, it takes somewhere between 9 and 20 positive comments to overtake that one negative comment. Online, the research is still a bit fuzzy, but it seems like that's been amplified to about 90. It's tough to tell when someone specifically doesn't buy as a result of a negative comment or a poor rating.
Not all ratings and negative comments are created equal. I recently wrote about how other factors can affect negative experiences. Since then, Ive had a few other experiences that have colored my suggestions on the matter. I needed to dispose of an old freezer. I know that freon is a bit of a bio hazard, so I was anticipating a little online research to figure out the best way to dispose of it safely. I searched for "appliances (myhometown) tx" and the result was (of course) a Google map with a list of locations near me. I'd heard of a few, but also noticed that there were 11 reviews for a company I'd seen, but never dealt with before. I read the first 5 star rating and felt good about it. I read the next listing- a one star listed and thought "what an idiot" I also kinda didn't believe that it was true. The next one was a 5 star rating and seemed genuine etc... The point is, people are not idiots and if you're going to rant and rave like a lunatic, it's unlikely you'll actually have a negative impact unless your comment drags down an average star rating. Seriously? Do you NEED to use all caps to get your point across? All Caps=lunatic in my book of online etiquette.
I got into an interesting dialogue with the folks over at advanced appliances and even had to convey the unhappy news that they lost their wonderful domain name and had a few negative comments. They were truly interested in repairing the damage that may have been done. In their case, my suggestion was to find the negative results, see if they can figure out who the customers were, and call and apologize and maybe even give them some money back or something like that. Importantly, I also suggested that they not ask them to remove the negative comment. I suggest not bringing it up at all. Just call because it's raining and you're in the habit of calling previous customers who had complained and you're catching up. the chance is good that they'll either try to remove their comment or may report that the company cared enough to call. Notice, I did NOT say they should lie about where they found the complaint. The customer will know- it's just more important to be genuine about your concern for the customer and their happiness. Everyone screws up sometimes. People realize that and I think it's far better to respond, attempt to repair and leave the negative comment alone.
Other negative experiences have a far better chance of damaging companies when the complaints are well thought out, true, and based on a situation that a customer service Representative has absolutely no ability to control. Small local companies can just be genuine, try harder and try to do more good than bad. Large companies are still not nearly careful enough about their business practices especially in a down economy. They simply haven't adapted and are wondering why their shareholder value has been eroding. I got an e-mail from Brinks Home Security a few days ago to announce their name change to Broadview Security. I can say honestly that I still don't really care why they changed their name. I read the e-mail and figured out it was as a result of what they called their "spin off last fall as a separate publicly traded company." OK really? I don't see any value there as a customer. I liked being a "Brinks" customer. That yard sign was the only reason I'd signed up with them in the first place. In fact, I realized that I hadn't actually been using my alarm very much. Actually- I had reviewed this bill last August and nearly cancelled then.
It may seem fairly obvious where this story is going, but lets just say that the "customer loyalty" guy I spoke to wasn't at all concerned that I don't appreciate a contract that automatically renews for a year at a time. You see- last August, I was thinking about completely eliminating my "home" phone and going cell phone only like most of my high tech friends. My security system was preventing me for doing that. When I explained the situation, the very nice customer loyalty person I'd spoken with told me she'd give me free service for three months while they finished working out their online monitoring technology. I thought that sounded pretty good. Then, in December or January, I realized that my bills had resumed and when I called to learn about this wonderful new technology, I was told it would cost me a bunch of money OR I could sign a new three year contract. Yikes. Come on people really? Apparently they just want customers regardless of the "cost." This very nice customer loyalty person tricked me into an automatic renewal for another year. I'm sure they would say that they were clear about it, but I'm STILL not interested in being their customer. In this case, it's probably a customer service issue. It would have been nice if he'd just said, "Oh wow I can see how it would seem that way. It looks like you got kinda hosed." I wouldn't have minded if he'd even said something like, "We're a pretty big company and sometimes our policies may seem a little strange, but it's a pretty common business practice." Pretty much anything that showed that he cared about the way I felt like I'd been treated.
I truly wish that companies were able to understand that their customers are people and people want to enjoy things, try new products, feel safe, find good services, and all the other things that go along with being a consumer. Most people I know don't want to be tricked or treated poorly. If you're concerned about what to do or what not to do regarding your online reputation, my best advice is to be human first. Then figure out ways to communicate that policy to your customers. It's just the right thing to do.
Online Reputation Management is the practice of understanding the buzz about a product, person, or even an offline company. An online reputation is just as often good as it is bad. In fact, true online reputation management is essentially public relations in the digital world of instant communications.
There are lots of similarities to the standard tenets of Public Relations, but many phenomenon have been amplified. "It's been said" and many have learned first hand that one positive comment made from one person to another may impact a sale in a positive way. Offline, that may translate to a 1 to 1 ratio. Online, if the information is "shared" through social media etc., it may create a 4:1 ratio (4 sales to every 1 positive comment) as a result of a "viral spread." The dangerous part is the amplification of the negative feedback loop.
In the offline world, the basic marketing principle is that for every one negative comment, it takes somewhere between 9 and 20 positive comments to overtake that one negative comment. Online, the research is still a bit fuzzy, but it seems like that's been amplified to about 90. It's tough to tell when someone specifically doesn't buy as a result of a negative comment or a poor rating.
Not all ratings and negative comments are created equal. I recently wrote about how other factors can affect negative experiences. Since then, Ive had a few other experiences that have colored my suggestions on the matter. I needed to dispose of an old freezer. I know that freon is a bit of a bio hazard, so I was anticipating a little online research to figure out the best way to dispose of it safely. I searched for "appliances (myhometown) tx" and the result was (of course) a Google map with a list of locations near me. I'd heard of a few, but also noticed that there were 11 reviews for a company I'd seen, but never dealt with before. I read the first 5 star rating and felt good about it. I read the next listing- a one star listed and thought "what an idiot" I also kinda didn't believe that it was true. The next one was a 5 star rating and seemed genuine etc... The point is, people are not idiots and if you're going to rant and rave like a lunatic, it's unlikely you'll actually have a negative impact unless your comment drags down an average star rating. Seriously? Do you NEED to use all caps to get your point across? All Caps=lunatic in my book of online etiquette.
I got into an interesting dialogue with the folks over at advanced appliances and even had to convey the unhappy news that they lost their wonderful domain name and had a few negative comments. They were truly interested in repairing the damage that may have been done. In their case, my suggestion was to find the negative results, see if they can figure out who the customers were, and call and apologize and maybe even give them some money back or something like that. Importantly, I also suggested that they not ask them to remove the negative comment. I suggest not bringing it up at all. Just call because it's raining and you're in the habit of calling previous customers who had complained and you're catching up. the chance is good that they'll either try to remove their comment or may report that the company cared enough to call. Notice, I did NOT say they should lie about where they found the complaint. The customer will know- it's just more important to be genuine about your concern for the customer and their happiness. Everyone screws up sometimes. People realize that and I think it's far better to respond, attempt to repair and leave the negative comment alone.
Other negative experiences have a far better chance of damaging companies when the complaints are well thought out, true, and based on a situation that a customer service Representative has absolutely no ability to control. Small local companies can just be genuine, try harder and try to do more good than bad. Large companies are still not nearly careful enough about their business practices especially in a down economy. They simply haven't adapted and are wondering why their shareholder value has been eroding. I got an e-mail from Brinks Home Security a few days ago to announce their name change to Broadview Security. I can say honestly that I still don't really care why they changed their name. I read the e-mail and figured out it was as a result of what they called their "spin off last fall as a separate publicly traded company." OK really? I don't see any value there as a customer. I liked being a "Brinks" customer. That yard sign was the only reason I'd signed up with them in the first place. In fact, I realized that I hadn't actually been using my alarm very much. Actually- I had reviewed this bill last August and nearly cancelled then.
It may seem fairly obvious where this story is going, but lets just say that the "customer loyalty" guy I spoke to wasn't at all concerned that I don't appreciate a contract that automatically renews for a year at a time. You see- last August, I was thinking about completely eliminating my "home" phone and going cell phone only like most of my high tech friends. My security system was preventing me for doing that. When I explained the situation, the very nice customer loyalty person I'd spoken with told me she'd give me free service for three months while they finished working out their online monitoring technology. I thought that sounded pretty good. Then, in December or January, I realized that my bills had resumed and when I called to learn about this wonderful new technology, I was told it would cost me a bunch of money OR I could sign a new three year contract. Yikes. Come on people really? Apparently they just want customers regardless of the "cost." This very nice customer loyalty person tricked me into an automatic renewal for another year. I'm sure they would say that they were clear about it, but I'm STILL not interested in being their customer. In this case, it's probably a customer service issue. It would have been nice if he'd just said, "Oh wow I can see how it would seem that way. It looks like you got kinda hosed." I wouldn't have minded if he'd even said something like, "We're a pretty big company and sometimes our policies may seem a little strange, but it's a pretty common business practice." Pretty much anything that showed that he cared about the way I felt like I'd been treated.
I truly wish that companies were able to understand that their customers are people and people want to enjoy things, try new products, feel safe, find good services, and all the other things that go along with being a consumer. Most people I know don't want to be tricked or treated poorly. If you're concerned about what to do or what not to do regarding your online reputation, my best advice is to be human first. Then figure out ways to communicate that policy to your customers. It's just the right thing to do.
Thursday, July 2, 2009
Action Oriented Metrics
Here's another post encouraging practicality as you endevour to measure your interactive activities. I've already written a few times about the benefits of finding and tracking your KPIs (Key Performance Indicators). I'd like to focus on the basic ideas behind KPIs and a little more about what they are nd what they are not.
Yesterday, I attended a lunch meeting hosted by the Houston Interactive Marketing Association. Scott Berg from HP gave a presentation titled The "Big Bang" of Consumer and Business Marketing Interestingly, there were alot of similarities between the concepts he was presenting and the table conversation prior to the presentation. Scott hit a few points over and over again. He emphasized how important search is and how it ties the rest of your online activities together and he said a few times that some metrics like "time on site" are not very valuable metrics to look at. I immediately realized that I'd spend about an hour on Verizon's site as I was trying to use it to download a new driver for my wireless card and then diagnose the trouble I was having. I truly would have preferred to have spent LESS time on their site the other day. If you are one that considers a long time on site to be a good KPI, I would encourage you to carefully consider that metric. Ironically, I've mentioned previously that Time on Site CAN be a good KPI. That's true- it may be a good KPI especially with a lack of other metrics.
Scott says that he has asked colleagues at HP to explain to him what he can learn from a number like time on site. In HP's case, it's possible that they may be interested in tracking time on site on their support pages in an attempt to reduce the average whereas they may want their time of site to remain stable on their product related pages.
However, if you are a publisher, time on site may be valuable to understand "engagement." If you are attempting to sell banner ads, you may be most interested in the time on site and the number of pages per visitor. You'll likely be interested in the ratio between the two and hope that they both grow over time.
The question still remains. What can you learn from a number like time on site? What can DO with that number? For that matter, what can you DO with any KPI? The idea of action oriented metrics is typically a tricky one. It requires exploring your data to look for changes that may be clues to things that are going well or things that are not going as well as they were or as well as you'd like.
KPI's are only as good as your ability to act on them. There are some numbers that will change over time inexplicably and there may be absolutely nothing you can do about it. That's very frustrating to a marketer. The big trick is to figure out the answer to that bigger question. What are my action oriented metrics? What can I DO with all these numbers?
I had the pleasure of working with lots of really smart people in my career. One of the smartest and most influential was certainly John Marshall the former CEO of ClickTracks (now Lyris). He was passionate about lots of things. Everyone in the company knew full well how much he disliked pie charts- especially 3d pie charts. We all also understood how much he disliked the concept of a dashboard. His point is very applicable to this topic. Consider the very thing that an analytics dashboard is typically modelled after. How often do you look at the dashboard of your car? When you first get your car, you need to figure out what each "metric" is and why it is valuable to you. Then- once you start driving, it's pretty typical to only glance down at it from time to time but especially when you see a cop. Perhaps you could liken that to a suprise visit from your boss who walks into your office to find out why your online sales are down. The fuel guage becomes important once the little yellow light comes on. The point is that some of the data provided on the dashboard of your car provide you with actionable data. You know exactly what to do with it when you see it. If you have an analytics dashboard, do you have any idea what to do with the data when it changes? Where is the accelerator for your website? Perhaps that's your paid search campaign that needs more budget or a tweek of the keywords you are buying. Maybe you need to consider a special offer to repeat customers along with a way to promote it to them? Or perhaps the number of visitors to your website means absolutely nothing to you and there is simply nothing that needs to be done about it.
Ok great. So now how do you find your action oriented metrics? I suggest you think like your customer. Go to your website as a potential customer. Have a few people who have never been to your site do the same thing. Does it work in the ways that you would expect? Perhaps you have a sales funnel that you can articulate in your analytics package. If you've recently changed anything, verify that it's working the way you expect. Think about your website and the experience your customer is likely to desire. To borrow a concept from Scott Berg, think of your site as an ecosystem. What metrics are important in YOUR ecosystem? Find your baseline and your benchmarks and you're likely to discover your action oriented metrics in that process.
Yesterday, I attended a lunch meeting hosted by the Houston Interactive Marketing Association. Scott Berg from HP gave a presentation titled The "Big Bang" of Consumer and Business Marketing Interestingly, there were alot of similarities between the concepts he was presenting and the table conversation prior to the presentation. Scott hit a few points over and over again. He emphasized how important search is and how it ties the rest of your online activities together and he said a few times that some metrics like "time on site" are not very valuable metrics to look at. I immediately realized that I'd spend about an hour on Verizon's site as I was trying to use it to download a new driver for my wireless card and then diagnose the trouble I was having. I truly would have preferred to have spent LESS time on their site the other day. If you are one that considers a long time on site to be a good KPI, I would encourage you to carefully consider that metric. Ironically, I've mentioned previously that Time on Site CAN be a good KPI. That's true- it may be a good KPI especially with a lack of other metrics.
Scott says that he has asked colleagues at HP to explain to him what he can learn from a number like time on site. In HP's case, it's possible that they may be interested in tracking time on site on their support pages in an attempt to reduce the average whereas they may want their time of site to remain stable on their product related pages.
However, if you are a publisher, time on site may be valuable to understand "engagement." If you are attempting to sell banner ads, you may be most interested in the time on site and the number of pages per visitor. You'll likely be interested in the ratio between the two and hope that they both grow over time.
The question still remains. What can you learn from a number like time on site? What can DO with that number? For that matter, what can you DO with any KPI? The idea of action oriented metrics is typically a tricky one. It requires exploring your data to look for changes that may be clues to things that are going well or things that are not going as well as they were or as well as you'd like.
KPI's are only as good as your ability to act on them. There are some numbers that will change over time inexplicably and there may be absolutely nothing you can do about it. That's very frustrating to a marketer. The big trick is to figure out the answer to that bigger question. What are my action oriented metrics? What can I DO with all these numbers?
I had the pleasure of working with lots of really smart people in my career. One of the smartest and most influential was certainly John Marshall the former CEO of ClickTracks (now Lyris). He was passionate about lots of things. Everyone in the company knew full well how much he disliked pie charts- especially 3d pie charts. We all also understood how much he disliked the concept of a dashboard. His point is very applicable to this topic. Consider the very thing that an analytics dashboard is typically modelled after. How often do you look at the dashboard of your car? When you first get your car, you need to figure out what each "metric" is and why it is valuable to you. Then- once you start driving, it's pretty typical to only glance down at it from time to time but especially when you see a cop. Perhaps you could liken that to a suprise visit from your boss who walks into your office to find out why your online sales are down. The fuel guage becomes important once the little yellow light comes on. The point is that some of the data provided on the dashboard of your car provide you with actionable data. You know exactly what to do with it when you see it. If you have an analytics dashboard, do you have any idea what to do with the data when it changes? Where is the accelerator for your website? Perhaps that's your paid search campaign that needs more budget or a tweek of the keywords you are buying. Maybe you need to consider a special offer to repeat customers along with a way to promote it to them? Or perhaps the number of visitors to your website means absolutely nothing to you and there is simply nothing that needs to be done about it.
Ok great. So now how do you find your action oriented metrics? I suggest you think like your customer. Go to your website as a potential customer. Have a few people who have never been to your site do the same thing. Does it work in the ways that you would expect? Perhaps you have a sales funnel that you can articulate in your analytics package. If you've recently changed anything, verify that it's working the way you expect. Think about your website and the experience your customer is likely to desire. To borrow a concept from Scott Berg, think of your site as an ecosystem. What metrics are important in YOUR ecosystem? Find your baseline and your benchmarks and you're likely to discover your action oriented metrics in that process.
Subscribe to:
Posts (Atom)