Making Web Analytics Actionable (300 words minimum)

JL12345

Web_Analytics_An_Hour_a_Day_----_Chapter_12_Month_7_Competitive_Intelligence_and_Web_2.0_Analytics.pdf

Home >Computer Science homework help >Making Web Analytics Actionable (300 words minimum)

297

■ M

O N

T H

7 : C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

Month 7: Competitive Intelligence and Web 2.0 Analytics

The Web is evolving so fast that what we had yes-

terday in terms of data is slightly less relevant

today. We just get a handle on static pages, and

here come dynamic sites. We get a handle on that,

and here comes personalization and targeting. We

go at it again and get a handle on that, and here

comes social media (blogs, RSS, user-generated

content). Then there are rich Internet applications

(RIAs) and even more new things.12

30652c12.qxd 5/1/07 12:48 PM Page 297

Kaushik, Avinash. Web Analytics : An Hour a Day, John Wiley & Sons, Incorporated, 2007. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=302273. Created from apus on 2020-09-07 20:35:44.

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

298

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

The challenge is not just keeping up with the breakneck pace of web evolution. It is that what we can measure to make our businesses successful is finite—so lots of change, not enough strategic things to measure. How do you survive?

This chapter covers two topics that will become critical in the near future: com- petitive intelligence (how to move outside your company data silo) and Web 2.0 analyt- ics (how to measure RIAs and Really Simple Syndication, or RSS, metrics). Your business likely doesn’t yet need these difficult and expensive analytics. But when you need to use them, you’ll be prepared.

Competitive Intelligence Analytics

In Chapter 2, “Data Collection—Importance and Options,” I explained why competi- tive intelligence is so important and the options available when it comes to accessing competitive data. To briefly recap, doing some level of competitive intelligence analysis is extremely important for the following reasons:

• Currently most reporting that you do probably uses your company data, and that presents a very compartmentalized view of reality. It is important to have the external context within which you can place your performance (even the context of your industry or your competitors).

• This analysis will tease out causality between your own company efforts and results.

• It will help you optimize your acquisition and marketing strategies based on what you can learn from your competitors.

• It will help you develop a competitive advantage.

Chapter 2 also discussed three options at your disposal from a data collection and analysis perspective, some free, some at cost. (Please see Chapter 2 for the benefits and pitfalls of each approach.)

Panel-based measurement: A group of recruits form a panel, their web behavior is monitored, and that data is analyzed.

ISP-based measurement: Data is collected from the various ISPs in the world, and this anonymous data is combined and provided for analysis.

Search engine data: Increasingly search engines such as Google and MSN are providing, for free, access to the massive data that they collect.

Each of these options has something of value that they can offer to help you do competitive analysis. Regardless of your company size or type of business, it is impor- tant to emphatically state that you can benefit from doing competitive analysis. If you are a Fortune 1000 company, get access to the paid competitive data providers. If you are a Fortune One Billion company, you can still use some of the free resources.

30652c12.qxd 5/1/07 12:48 PM Page 298

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

299

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

In the next two weeks, you will dive deep into three specific sets of options that will help you move your competitive intelligence offering from 0 to 900 miles per hour—very fast pace, very practical advice, very fast insights. I promise you, this is a lot of fun, it really is. Put on your Sherlock Holmes hat.

Week 1: Competitive Traffic Reports

The first thing you probably want to know is how many visits you get as compared to your competitor! You have a couple of options for running these reports.

Monday: Share of Visits by Industry Segment

By using tools from companies such as Hitwise (or comScore) for an industry segment (for example, Software—Technology), you have the ability to measure the share of traffic that you are getting and compare that to your competitors’ share. It should show you at a quick glance the big boys and girls in your segment whom you are fighting with.

You not only can see how you are performing in terms of your industry segment (Software in Figure 12.1), but you also have the ability to create your own custom competitor set and compare your traffic statistics to those of your competitors.

Figure 12.1 Market share by industry segment

30652c12.qxd 5/1/07 12:48 PM Page 299

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

300

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

You can input your core competitors into the tool and measure the share of vis- its that you have on your own turf (Figure 12.2). This is where you’ll see whether you are really dominating or your competitors are kicking your butt. On the other hand, the benefit of Figure 12.1 is that it often highlights your “unknown” competition that has the greatest share in terms of visits. (Because there is only a finite set of visits, if they win too much, you lose, even if you don’t directly compete with them.)

Figure 12.2 Market share by custom competitor segment

As always, it is optimal to watch trends over time to get a real good feel for how you are doing (Figure 12.3).

Figure 12.3 Market share trends over time

It is important to measure your share of traffic because over time you’ll be able to accurately identify whether the massive traffic increases that you see in your Omni- ture, WebTrends, WebSideStory, or Coremetrics data are a result of something you have done or general industry trends. Often it can explain increases in traffic on your site when you have not done anything (your competitor is running a massive campaign and you are simply getting a “halo effect” from that, for example).

Tuesday: Upstream and Downstream Traffic against Competition

Referring URLs in your web analytics tool’s report give you an idea of which websites are referring traffic to your website. But usually 50 percent of referrers (those who are sending you traffic) are blank. By accessing this data in a competitive intelligence tool,

30652c12.qxd 5/1/07 12:48 PM Page 300

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

301

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

you can get complete access to your referrers. More important—and how cool is this—you can also report on where visitors go next after visiting your website (data you have simply no access to in your clickstream applications unless you link to other sites). Figure 12.4 shows the report for www.peachtree.com and it illustrates, on the left, which websites are sending traffic to www.peachtree.com and, on the right, where visi- tors from www.peachtree.com go next.

Figure 12.4 Upstream and downstream traffic analysis

This is a great way to peek under the covers and see who is sending traffic to your competitors. You can then compare that to your own strategies and perhaps iden- tify future business partners for yourself. This report also gives you a great feel for mindset as you look at where people come from and then what site they go to next. If you have a small cluster of sites that people see before they come to you, you can make an inference about customer intent. For example, in Figure 12.4 it may be great for peachtree.com that such a large percent of traffic goes to subscription next (perhaps that is how Peachtree’s owner, Sage Software, makes money) and then on to Staples.com (which is where visitors can buy Peachtree products, maybe because those products are not sold at peachtree.com).

Wednesday and Thursday: Competitor Traffic by Media Mix

Media mix is defined as the core streams of traffic to your site: direct marketing, affili- ate, search, and direct/other. This report will show your media mix as compared to your competition, as you can see in Figure 12.5.

30652c12.qxd 5/1/07 12:48 PM Page 301

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

302

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Figure 12.5 Competitor media mix analysis—search engine, email (DM), and

direct traffic

This report is insightful because it provides a peek into the acquisition strategy of your competitor, something you would have a hard time getting otherwise. For example, by looking at the media mix report in Figure 12.5, you can summarize the percent of traffic that this website is getting as a result of email campaigns and search engines (including campaigns). You can see exactly when the spending on these cam- paigns is up (as shown by the graph going up or down) and where you might have opportunities if you wanted to avoid going head-to-head with your main competitor.

Is the media mix for your competitors the same as yours? Very different? Should it be the same? What are their core affiliate sites that are driving traffic to them? Should you go after them as well and have a relationship? Perhaps an alternative affiliate net- work? How about direct marketing, now that you know how efficiently (or not) they are using email, what should be your strategy?

These are complex questions that you can’t even begin to answer with the options you have available through standard web analytics tools. But with competitive intelligence you can start that journey. Even adoption of one strategy based on this data could be worth hundreds of thousands of dollars. So it is complex, and it also costs money to get access to this data, but the payoff could be huge. As always, you have to assess in your own company whether you can take action on this type of data.

Friday: Alexa Daily Reach Reports

Alexa Internet is always greeted with howls of protest, and for some good reasons. Alexa collects data from several million users who have installed the Alexa toolbar in their browsers. It has a bias toward Windows users, who use Internet Explorer. Ratings in Alexa higher than a rank of 100,000 are not reliable. (Please read information at http://snipurl.com/alexaA and http://snipurl.com/alexaB to learn more.)

30652c12.qxd 5/1/07 12:48 PM Page 302

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

303

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

In spite of that, if you want to see traffic trends as compared to those of your competitors—while holding all the bias as equal for you and your competitors—Alexa can be an acceptable resource. This is true especially if you observe trends over time, as shown in Figure 12.6.

Figure 12.6 Daily Reach analysis from Alexa

This graph is for the major sites in the world. However, you can imagine how you could glean insights about your success as compared to your competitors if you had this type of graph. The growth that Google has shown is impressive, as is that of Windows Live (www.live.com), which came from nowhere and continues to grow (with- out cannibalizing MSN either).

You can do the preceding type of analysis for smaller sites as well (as long as their individual ranks are under 100,000). Figure 12.7 shows analysis for web analytics bloggers.

Figure 12.7 Daily Reach analysis, for web analytics bloggers, from Alexa

30652c12.qxd 5/1/07 12:48 PM Page 303

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

304

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Figure 12.7 shows a three-month trend for web analytics bloggers. Although their individual web analytics metrics might not be available, you can glean insights from their trends in the context of each other.

It is important to stress that you should not look at the absolute numbers on Alexa. You should look only at trends. Simply ignore the y-axis of that graph. If you want to get a directional read, for free, about the trends of your website and your com- petitors’ websites, this is a good resource.

Week 2: Search Engine Reports

I have said at various times in this book that search engines are prominent influences of customer behavior. For this reason I have also stressed the importance of doing deep search engine optimization and pay per click analytics (see Chapter 8, “Month 3: Search Analytics—Internal Search, SEO, and PPC”). Those are beneficial and necessary, but for you to get a leg up in this dog-eat-dog search world, you need to get into com- petitive intelligence!

Monday: Share of Search and Search Keywords

You will start the week by gaining a solid understanding of how you are performing in the highly competitive world of capturing traffic from search engines. You’ll also focus on which search keywords are performing well for you and your competitors.

Share of Search Report

If 80 percent of the traffic on the Web starts on a search engine, how much share do you have of that traffic? What about compared to your arch nemesis? The Share of Search report will answer these and other questions. See Figure 12.8.

All the competitors being compared in Figure 12.8 are outperforming their industry average (the lowest line), but two of them are doing amazingly better. If you are not one of them, this report can quickly help you highlight why you might not be getting as much traffic from search engines (something you won’t understand from your clickstream data) or that your SEO and PPC efforts are coming to nothing and need to be retooled.

You also have an ability to drill down one more level in this report and compare your share of search engines against your competitors, as shown in Figure 12.9.

T i p : Don’t look at your site’s Alexa ranks (you might not get the right bearing and therefore might misin-

terpret the data). In addition, remember to compare like-minded sites only (to keep the bias of the data col-

lection methodology as neutral as possible).

30652c12.qxd 5/1/07 12:48 PM Page 304

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

305

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

Figure 12.8 Share of Search competitive report

Figure 12.9 Search Engine Share report for Pepsi and Coke

In Figure 12.9, you can see the share of search engine data for Pepsi and Coke. Let’s say you work for Pepsi. You can quickly see that you are not performing very well against Coke when it comes to the most dominant search engine in the world (Google). Unless Pepsi deliberately focuses on other search engines (a valid strategy, of course), Pepsi might need to do a quick investigation of whether their site is indexed correctly in Google. (There could be a whole host of other possibilities, but this report will at least highlight an important concern.)

Share of Brand and Category Keywords

Your clickstream analytics can help you highlight what keywords or key phrases are your top 10 or 20. In reality, a whole bunch of traffic for most websites comes from the top 20 or so key phrases. Getting access to competitive intelligence data helps you understand what is your share of clicks is from search engine results—for those key- words when people search at Google or Yahoo! or MSN.

You can start with a look at your branded key terms to understand your share of clicks. It is entirely possible that you dominate those clicks (simply because it is your brand and you are probably optimized for your own brand terms). Sometimes this report can highlight others who might be doing illegal things to steal your traffic from the search engine.

30652c12.qxd 5/1/07 12:48 PM Page 305

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

306

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

But where this report really comes in handy is when you search for category key phrases (phrases that are not tied to your brand—for example, soft drinks instead of Pepsi). Category key phrases are more important for businesses because they are the source of prospects, people who might not yet know your brand and therefore are potentially not your customers. If you are looking to grow your customer base, you really want to focus on category terms. Figure 12.10 shows a category key phrase report.

Figure 12.10 Report for share of category key phrase payroll software

It is easy to look at Figure 12.10 and understand precisely who is capturing all the prospects looking to buy payroll software for their businesses. If you are 2020software.com, cyma.com, sagemas.com, or realtaxtools.com, you should really be worried that you don’t even show up in the top 10 (and your only way to capture traffic is to do PPC campaigns, which are an expensive way to get traffic). If you are paycycle.com, you should be happy because you are doing so well in this competitive intelligence report.

Everyone wants to capture prospects. Simple reports like these can fundamen- tally change the direction and emphasis of your search marketing efforts, if not your fundamental acquisition strategy. Demand this kind of report from wherever you get your competitive intelligence data.

Tuesday: Search Keyword Funnels and Keyword Forecasts

This day’s task is kind of cool. You will learn how to get insight into the shopping / con- sideration behavior of your customers by getting a peek into their behavior on a search engine and also use data from the search engine (available for free) to understand how the search engine is expecting demand for keywords to change over the coming months.

30652c12.qxd 5/1/07 12:48 PM Page 306

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

307

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

Search Funnel Report

Microsoft adCenter Labs provides access to this wonderful little report for free (http://snipurl.com/msnfunnel). The search funnel report helps you understand cus- tomer intent by reporting what people search for before they search for your top key phrases and what they look for after. See Figure 12.11.

Figure 12.11 Search funnel report for the key phrase peachtree

The pre-funnel (top portion of the figure) provides insights about what is on people’s minds before they think of you. It outlines what your competitors are feed- ing you as well as what nonbranded key phrases are most relevant. It also includes a couple of surprises—for example, it is hard to discern the relationship between pella and peachtree, but if you are Peachtree, it might be worth investigating.

Ditto for the post-funnel (bottom portion of the figure). Of course, after people search for Peachtree, they want to go back and search for QuickBooks (a direct com- petitor). But it is surprising that they go back to look for peachtree business products, peach tree accounting, peachtree software, and peachtree accounting. If the results of the search for Peachtree (which is a branded term for Sane Solutions and hence easy to “capture”) are optimized, the web traffic should be landing on pages that are opti- mized to give answers related to all four key phrases. The traffic should not have to go back and search for that detail again.

You can gain some very actionable insights from an easy-to-use free tool (even if the data is from searches done only on MSN or Windows Live).

Keyword or Key Phrase Forecast

Your company probably has internal plans for pay per click campaigns around your product launches or selling seasons. What this wonderful free tool, MSN adCenter Keyword Forecast (http://snipurl.com/msnforecast), allows you to do is to get an out- sider’s opinion of what the next few months look like for your top 10 keywords. The tool will compare your forecast to your competitors as well, as shown in Figure 12.12.

30652c12.qxd 5/1/07 12:48 PM Page 307

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

308

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Figure 12.12 Keyword forecast report from MSN adCenter

It is to be expected that during the holiday season, sales of digital cameras will pick up. But the folks running the Nikon and Olympus web strategies don’t know that Canon is doing so much better than them, and is predicted to do much better than them even during the holiday season. Could they use this data to adapt their strategies to do much better against Canon? They could come up with a more robust strategy around nonbranded keywords to counter this strong brand keyword trend.

If only for your MSN advertising strategies, you can consider the keyword fore- cast report as valuable input. You can also determine the size of the opportunity better for your company.

Wednesday: Keyword Expansion Tool

If you remotely touch search marketing and have the tough job of generating an opti- mal set of keywords to bid on, you realize that the competition is really hard out there, especially for the keywords that you can think of or get from your web analytics appli- cation. How do you expand your keyword list to find words or phrases that you don’t know? How do you start to look for opportunities for arbitrage (profiting from differ- ences between supply and demand)?

Google has a great free keyword expansion tool (http://snipurl.com/ googexpansion) that you can use to quickly expand your keyword list and build up your very own long tail (keywords and phrases that are specific and yield few searchers but in total could yield more than your main keywords).

It works quite simply. You go to the website and type in some keywords you are interested in (for the example in Figure 12.13, I used dell, laptops, dimension, and xps).

30652c12.qxd 5/1/07 12:48 PM Page 308

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

309

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

Figure 12.13 Keyword expansion tool from Google

After you click Get More Keywords, the tool provides a list of additional key- words that you could consider bidding on to build a robust portfolio of keywords (182 additional keywords for my four in December 2006). Not too shabby for five seconds of work. Just try it with your top four, five, or fifty keywords, and you’ll be surprised. You may also get more praise from executives because you are not just reporting things from your analytics tool.

The tool also has a couple of other great features. Figure 12.14 shows a report from the keyword expansion tool that will tell you how much competitive bidding can be expected for each key phrase and how much inventory (search volume) is predicted to be available for each key phrase. This information can go directly into influencing your bidding strategy.

Figure 12.14 Keyword expansion tool competition and volume

Additionally, you can find opportunities for arbitrage where there is not really a lot of competition but lots of searchers (for example, xps motherboard in Figure 12.15). There are no hard numbers in either column, but Google has at least gotten you started on a path with some guidance (rather than you going in blind).

It is quite likely that if you are a large spender of PPC dollars, you have an agency that can help you expand your keywords in much more sophisticated ways. They probably can also find arbitrage opportunities with some kind of advanced algorithm. But if you are not a large spender, you can use this tool. Even if you are a large spender’s analyst, now you have a smart way to double-check what your agency is doing.

30652c12.qxd 5/1/07 12:48 PM Page 309

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

310

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Figure 12.15 Keyword expansion tool from Google

Thursday and Friday: Demographic and Psychographic Reports

Few website owners have any awareness of the demographic nature of their visitors. Yet knowing whether your visitors are from Mars or Venus could be a major influencing factor in your website design and experience. Access to demographic and psychographic data (customer segments and associated attributes related to personality, values, atti- tudes, interests, or lifestyles) can also be valuable in helping you find new sources of reaching your targeted segment. It can help you optimize your campaigns and more.

Demographic Prediction

You can use the free demographic prediction tool at MSN adCenter (http://snipurl .com/msndemo) to understand the basic demographic information about your websites (or those of your competitors). See Figure 12.16.

Figure 12.16 Demographic prediction for pepsi.com

The more popular your website, the more likely that the data is accurate. It is sourced from all data that MSN has about its searchers. You can use this data to opti- mize your website experiences or to validate that your traffic acquisition strategies are working as they were intended to. It can help have some influence on your product marketing, customer experience, and customer retention strategies.

30652c12.qxd 5/1/07 12:48 PM Page 310

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

311

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

You can also compare your demographic information to your competition’s. For example, coca-cola.com (on Dec 31, 2006) had a ratio of 0.24 male and 0.76 female, drastically different from Pepsi. Only someone at Pepsi can decide whether this is good or bad for them (and then figure out how to leverage this newfound information).

Psychographic Reporting

Perhaps the best use of demographic or psychographic data (the latter leverages the standardized Prizm lifestyles data in Hitwise) is to optimize your acquisition strategies and understand the influence of your competitors on the customer segments that you are interested in. Claritas offers Prizm customer segmentation and profiles that identify segments of customers in the United States (by geographic location, for example) and provide insights into their shopping habits, their demographic profiles, and other mar- keting information.

For example, you can use the tool in Hitwise (Figure 12.17) to run queries against their competitive intelligence data and not only identify the demographic pro- files of customers who visit your website, or your competitor’s website, but also under- stand the behavior of these targeted customer segments. Where do “Midlife Success” or “Movers & Shakers” or “Young Digerati” go when they are surfing? What are their preferred websites? You can optimize your campaigns accordingly.

Figure 12.17 Demographic and psychographic analysis in Hitwise

30652c12.qxd 5/1/07 12:48 PM Page 311

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

312

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

You can also get a lot more sophisticated by targeting people only in certain states, age groups, or household incomes. You can imagine how quickly you can extract maximum ROI from your online campaigns.

Competitive intelligence should be the cornerstone of your web analytics strategy if you are to move beyond your own company data silo and derive maximum actionability that adds to your company’s bottom line. The current crop of sources (including Hitwise, MSN, Google, and Alexa)—some free and some paid—each have their strengths and weakness (see Chapter 2). But if you familiarize yourself with them and then use the tools, each can help you become a little more knowledgeable and help move your com- pany ahead from a strategic perspective.

Two Competitive Intelligence Pitfalls to Avoid

You’ve learned why to do competitive intelligence analysis, the specific reports you can use, and

the specific actions you can induce as a result of your analysis. Here I want to share two types of

analyses that people jump into first when they hear of competitive intelligence. Both are usually

misleading, take up a lot of time, and in the end provide few actionable insights.

Conversion Rate

The instant tendency for anyone, especially senior management, is to ask for the competitor’s con-

version rate and compare it to their own. Usually this is a huge waste of time.

Even companies who are in exactly the same business have radically different business strategies

when it comes to the Web. For example, you could be driving all the sales via the web channel to the

detriment of other channels while your competitor could have a more holistic web, retail, and phone

strategy. So if you compare conversion rates, you are really comparing apples and mosquitoes.

For example, consider Best Buy and Circuit City, two electronics powerhouses. On the surface it might

seem like they are in the same business on the Web. But in reality, whether in their retail stores or on

their websites, they execute such radically different strategies that even if you knew what their web

conversion rates were, it would give you very little insight that you could exploit to your advantage.

It is a classic So What situation (see Chapter 5, “Web Analytics Fundamentals”). Let’s say your con-

version rate is 90 percent and your competitor’s is 9 percent. What would you do? What if it were

vice versa? Would you change your fundamental business strategy? Would it really matter?

N o t e : To leverage this data, you have to have a good understanding of who your customers are and what

customer segments fit with your company’s strategic objectives. Not every company has that understanding.

But if your company does, you need access to this data!

30652c12.qxd 5/1/07 12:48 PM Page 312

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

313

■ C

O M

P E

T IT

IV E

IN T

E L

L IG

E N

C E

A N

A L

Y T

IC S

Two Competitive Intelligence Pitfalls to Avoid (Continued)

There is also wide variance in how each competitive intelligence provider measures conversion.

Some will simply detect the presence of an HTTPS entry in the visitor session and assume that

moving to a secure area of the site is conversion. Others might try to pick up URL stem values that

could indicate conversion. Please be cautious of exactly how your competitive conversion numbers

are being provided (assuming that you have validated that your business model, business strategy,

and site structure is exactly the same as your competitor first).

Measuring simply to know rather than to take action is an expensive distraction that often pro-

vides a misleading sense of confidence (or panic, as the case may be).

Pages, or Content, Viewed

This one is tricky. It seems logical that you would want to know the pages that visitors are viewing

on your website and on your competitor’s website. A common question asked could be, How many

visitors are viewing the product detail pages on their site compared to ours? The following discus-

sion outlines some reasons why this could be a suboptimal use of time.

Any data you need for this comparison will need to be deeply customized because, sadly, no two

sites follow the same structure for content (and it changes all the time). As an example, here are

two pages for the exact same product:

http://www.circuitcity.com/ssm/Sony-Cyber-shot-DSC-H2-Digital-Camera/sem/

rpsm/oid/149326/catOid/-13062/rpem/ccd/productDetail.do

http://www.bestbuy.com/site/olspage.jsp?skuId=7698896&type=product&id=

1138084657346

In order for your reporting to be accurate, you would have to put in massive effort to ensure that

what you are calling Product Page on one site is the same as Product Page on the other site

(please see Chapter 6, “Month 1: Diving Deep into Core Web Analytics Concepts”). You can see how

this can get expensive and out of hand.

Attributing intent to a page view is quite a stretch. If you see this and the purpose of this page is

that, you must be trying to do z. This throws a kink in the content viewed analysis, especially across

different websites (and you would have to keep pace with all the changes on sites).

Imagine all the challenges you could have to overcome to measure page views (or content con-

sumed) on a site that is using RIAs and Flash and Ajax, where the total number of page views

could be 1 to any current competitive intelligence tool, when in reality the customer could have

“viewed” 50 “pages.”

This is another example of the So What situation. Apply the “so what” test and only after a satisfac-

tory answer to the test should you plunge into this analysis.

30652c12.qxd 5/1/07 12:48 PM Page 313

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

314

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

In summary, business models and strategy, acquisition campaigns, site structure/ experience, and customer mindsets and intent all influence the success of a competitor. As you do competitive intelligence, attempt to focus on core areas that will help you improve in areas where there is a closer apple-to-apple comparison of the data rather than where your differences could cause you to be misled. The report examples that I have provided are great places to start to look for competitive insights from which you can gain actual competitive advantage. It is great to get into this space now with lots of new companies with new data capture models being formed recently (Compete and Quantcast, to name a couple).

Web 2.0 Analytics

Web 2.0 is a much-brandished buzzword that means almost nothing or almost everything depending on who you talk to. To provide a definition, I’ll turn to the wonderful source of wisdom that is Wikipedia (http://snipurl.com/wiki20) and the excellent O’Reilly Sep- tember 2005 article called “What Is Web 2.0” (http://snipurl.com/oreilly20):

The phrase Web 2.0 refers to one or more of the following:

• The transition of websites from isolated information silos to sources of content and functionality, thus becoming computing platforms serving web applications to end users

• A social phenomenon embracing an approach to generating and distributing web content itself, characterized by open communication, decentralization of author- ity, freedom to share and reuse, and “the market as a conversation”

• More-organized and categorized content, with a far more developed deep-linking web architecture than in the past

• A shift in economic value of the Web, possibly surpassing that of the dot-com boom of the late 1990s

• A marketing term used to differentiate new web-based firms from those of the dot-com boom, which subsequently appeared discredited (because of the bust)

• The resurgence of excitement around the implications of innovative web applica- tions and services that gained a lot of momentum around mid-2005 In my humble opinion, this perhaps best captures the essence of Web 2.0.

You can imagine how each of the preceding bullets can cause challenges for web analytics practitioners when it comes to measurement and analysis. Our current sets of tools are not very prepared for a world beyond page view website architecture. Increas- ingly in the Web 2.0 world, it is harder to measure intent, and the click loses its value. One of the emerging trends of Web 2.0 is that increasingly content creators are losing control of their content and the process of creating customer experiences. Increasingly, customers are creating their own customized customer experiences (for the simplest example, think of mash-ups) and customers are being influenced more and more by social media.

30652c12.qxd 5/1/07 12:48 PM Page 314

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

315

■ W

E B

2 .0

A N

A L

Y T

IC S

Measurement systems for Web 2.0 will continue to emerge and evolve over time. There is not a lot of standardization at the moment, and it is hard to see where all this will end up. For this reason, I will cover two technologies and customer experiences that are very much Web 2.0 and the options that are available to you when it comes to measuring their success: RIAs and RSS. My hope is to share a new way of thinking for the new world and prepare you just a little bit more as it evolves around us.

Week 3: Measuring the Success of Rich Interactive Applications (RIAs)

Web experience for the past 10-odd years has meant HTML and just HTML on a page. There have been incremental improvements; many sites are now dynamic and can react to a trigger (a cookie, parameter, ID, and so forth) to present different content on a page. But essentially it is still a page. We have had Flash for a few years; think of demos or pop-ups displaying software or product features. We have had the ability to rotate the model of the camera 360 degrees before we decide to buy it.

In the last couple of years (primarily since 2005), we have truly stepped into a world where we are moving beyond product demos as “rich” experiences to fully func- tional desktop-application-type interactivity and experiences in an Internet browser. These RIAs are driven by technologies such as Asynchronous JavaScript and XML (Ajax), Adobe Flash, Adobe Flex, OpenLaszlo, and others.

A couple of famous examples of Ajax-based RIAs are Google Maps and Google Mail (Gmail). Both use Ajax, allowing a web page to be more interactive without hav- ing to refresh the page to serve up new content. Your entire experience in Gmail uses one URL (http://mail.google.com/mail). As you reply and delete and forward, the URL never changes. Often the page itself does not reload.

An example of a RIA-driven site is www.miniusa.com. You can go to the site and browse around and you’ll experience a rich media experience (in this case driven by Flash) that is like nothing you have seen before. Watch the URL and its changes to see how they are collecting data (or use View Source). You can view a Flex-driven sample website by visiting http://snipurl.com/flexstore. Click on the Products tab and try the experience.

Monday: Understand the Reality of RIAs

RIAs are still very new, as of early 2007, and you might not run into them in the course of your typical web surfing. Yet RIAs hold a lot of promise in terms of benefits to our customers. As always, such interactions come with their own set of challenges, and because we are still at the very early stages, there is the reality of where we are to deal with.

Beyond simply being cool, RIAs provide a number of benefits in terms of improved customer experiences on the Web. They help users find and display content

30652c12.qxd 5/1/07 12:48 PM Page 315

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

316

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

without waiting for pages to refresh (from the server). They are especially beneficial in improving multipage structured processes such as sign-ups for leads, or the cart and checkout process, or booking a hotel by selecting dates and rooms and credit card numbers, all on one page. Figure 12.18 shows a rich interactive experience, purchasing clothing, on gap.com, and more specifically how the site reacts, dynamically, when the add to bag button is clicked before the size of the jacket is chosen.

Figure 12.18 Gap.com RIA experience: error handling

RIA sites have much better error messaging on the page (so you don’t have to wait to click Submit to know you are missing a digit on your telephone number, for example). Perhaps one of my personal favorites is the capability to undo your last action. For example, let’s say you mistakenly deleted a product from your cart. Now, rather than going back to the product page and finding the item and adding it to the cart, you can simply click the Undo button.

Every web analytics vendor relies on a page and a “page view” to identify a dis- crete event. As you learned in Chapter 6, the existence of data in the URL or URL parameters identifies an event to the web analytics application. With some vendors, you can also pass a variable via JavaScript to indicate an action (“someone viewed a page in the Products group”).

The challenge with RIAs (Ajax or Flex) is that there is no such concept as a page. In Gmail, your URL does not change after you click Reply. For your entire Gmail experience, you’ll be on the same URL although you have seen 20 “pages” (a better way to think of this is that you would have initiated 20 actions or events). If you use

30652c12.qxd 5/1/07 12:48 PM Page 316

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

317

■ W

E B

2 .0

A N

A L

Y T

IC S

Flash, the browser just sees something.com/ria.swf (or similar) and a person could have interacted with 100 “pages.” But to the analytics tool, it is just one page view because it needs the URL or a combination of URL and parameters to provide your standard- ized reports.

The combination of the current architectures of the web analytics tracking tools and the radically different architectures of the new emerging tools such as RIAs makes it difficult to track anything out of the box except that someone loaded the RIA in their browser.

You’ll spend Tuesday and Wednesday learning about a couple of ways in which RIA experiences are being tracked. Still, it is extremely important to be aware that today we are at the dinosaur stage of evolution. There are no current standards, just a few vendors reacting to RIAs. At best what we are doing is reacting to the tracking technologies that we know of today.

As RIAs move beyond cool (and I predict that they will—just check out a demo of Adobe Apollo), new technologies will emerge. There will be standards, and it will be easier to implement tracking. We will make it from dinosaurs to humans, but it will take time and it will be radically different from the current “hacks” that are being used to measure RIA success. It is important to bear in mind that what exists today will go away and be replaced with something better.

Tuesday and Wednesday: Learn about Emerging “Standards” for RIA Tracking

Analysts are partnering with business and technical folks to transform thinking from page views to events. There are a few ways to do this. Each RIA is a piece of software with which our users interact, and each button click, choice selection, and mouse movement is a business event. Think of the action of Add to Cart or Insert Image (to create a cartoon) or Next or Update. In RIAs, each of these actions is akin to a click and refresh in the HTML world, only in the RIA no refresh is required. As you interact with Google Earth, these events are pan, drill down, move left, and so forth.

The first step in tracking RIAs is to identify why the RIA exists, what customer problems it is trying to solve, and then break down the experience into a series of events. After the business events are identified, there are a couple of different methods used to collect data.

The de facto standard is to extend the current JavaScript tags that you use on your website from your vendor (HBX, WebTrends, ClickTracks, IndexTools). Simply use a customized JavaScript tag and embed that into your business events so that if those events occur, the JavaScript tag will be fired off and a log will store the data.

The downside of using this method is that it relies on the many limitations of page views outlined earlier because essentially you are not capturing events as much as page views. This is suboptimal, but, for example, if you need to know only discrete

30652c12.qxd 5/1/07 12:48 PM Page 317

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

318

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

isolated things, this is fine. If your vendor allows you to pass more discrete variables (for example, the “calls to action” from the RIA), this method can be helpful, espe- cially if your vendor allows you to capture that data and do some custom processing to get the reports out.

Another option is to use custom-defined data collection methods to capture the data. You can leverage your web commerce platform’s native ability to capture events (ATG, for example, has this). Simply put, in the RIA experiences you embed the custom code that will capture not just the occurrence of the business event (say, Add to Cart) but also all the context around that event (quantity, product name, type, page, and so forth). Because we are leveraging an existing feature set from your web e-commerce platform, this development is extremely cost-effective and adds only an additional 20 percent to the developer time.

Omniture has recently introduced a solution called ActionSource that provides a customized way to collect the click activity in your RIA. Like the event-logging mecha- nism, it does not use JavaScript. But unlike the event-logging method, ActionSource provides a standard set of instructions and tools that your Flash developers can incor- porate into their development process to collect data, and your team does not have to do any custom in-house development.

Both of these methodologies allow you to collect a lot of data from the RIA but also ensure that you can tie the RIA data to other data sources—for example, your orders data that sits in Oracle or your customer satisfaction survey data.

Thursday and Friday: Learn the Steps for Successful RIA Tracking

We are in the very early phases of moving web experiences from HTML pages to rich interactive experiences. There are few standard tools that offer RIA (Flash, AJAX, others) measurement out of the box, so you will have to improvise a little bit. Regard- less of whether your tool supports RIA measurement, you will typically follow these recommended steps:

1. Partner with the business users to identify the core purpose of the RIA. What is the reason for the RIA’s existence? What problem are we solving for the cus- tomer? What are the “actions” that a customer will typically take? What are the success metrics?

2. After success metrics have been identified, step through the RIA experience with your business and technical counterparts and identify the key customer events that we will need to “tag” in order to capture that event being executed by our web customers.

3. Ask your technical folks to ensure that each identified event is tagged with the right piece of JavaScript tag or web server event-logging technology (for

30652c12.qxd 5/1/07 12:48 PM Page 318

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

319

■ W

E B

2 .0

A N

A L

Y T

IC S

example, if you are not ATG). Ensure that you are capturing not only the event but also some context, such as session_id, date/timestamp, perhaps that the per- son was adding Quicken Deluxe to the cart or in Google Earth clicked Pan on the city that they were looking at. It is important to QA this step, both so that all the events are tracked and you are collecting the right data.

4. Develop a simple Perl script, or Extract Transform Load (ETL) program load, to tag data into a simple database. You can use MySQL or Oracle or whatever database that you use in your company. If you already have a data warehouse for your web information, simply load the RIA data into that data warehouse. If you use a tool such as NetTracker, it will parse the logs and provide the data in a database-friendly format.

5. Use straight SQL, or a standard On-Line Analytical Processing (OLAP) tool if one is available, to run queries against the data to compute the success metrics.

If your web analytics vendor will allow you to capture events by passing variables via JavaScript tags, you can also use the functionality they have built into the web analyt- ics reporting tool to measure your success metrics (steps 4 and 5 would be different for you). Unica’s NetInsight is one such web analytics tool that has great capabilities built in.

In summary, although RIAs will bring richer desktop-type applications to your website browsers, it is important to realize that all the fluidity in the experience makes tracking harder, especially with current measurement systems. In terms of data collec- tion, there is also a paradigm shift. Currently we launch websites with some standard tags and they collect almost all the data you need to report on. After the fact, you can improve the data collection tag if you want but the tag is pretty much always there. With RIAs you have to put in a lot of effort up front to understand the business pur- pose and then identify the business events and ensure that they are tagged by develop- ers up front. If that front loading is not done, you have no data. It is as simple as that.

T i p : It is extremely important to test RIAs against currently standard web experiences (“dumb” HTML). In

many cases, you will find that simply testing in a lab usability environment will not quite represent how real

customers would react. Hence it is important to launch your RIAs in an A/B test environment to validate that

indeed it is an improvement as compared to the current experience. It is not uncommon for customers to

reject what they are not used to. It does not mean that you fall into despair and do not try; it means that you

test and you optimize the experience based on customer feedback. (Chapter 10,“Month 5: Website Experimen-

tation and Testing—Shifting the Power to Customers and Achieving Significant Outcomes,” covered experi-

mentation and testing in detail. Please review that chapter for details on how to go about doing this.)

30652c12.qxd 5/1/07 12:48 PM Page 319

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

320

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Week 4: Measuring the Success of RSS

From Wikipedia: RSS is a family of web feed formats used to publish frequently updated digital content, such as blogs, news, feeds or podcasts. Users of RSS content use programs called feed “readers” or “aggregators”: the user “subscribes” to a feed by supplying to their reader a link to the feed; the reader can then check the user’s sub- scribed feeds to see if any of those feeds have new content since the last time it checked, and if so, retrieve that content and present it to the user.

In other words, RSS is an absolutely wonderful method for your customers to sign up to get the content they are interested in from your website, but have it deliv- ered to them via a method and in a format and at a destination of their choosing.

Monday: Understand the Reality of RSS

RSS is changing the way your customers consume content. Its increased popularity means that web analytics applications, as they exist today in early 2007, are unable to accurately report data for your website.

Here is a quick example. I like keeping up with cricket news and I would typi- cally visit the BBC Sport website and read the news about the Indian cricket team, as shown in Figure 12.19.

Figure 12.19 BBC Sport: Cricket in India

30652c12.qxd 5/1/07 12:48 PM Page 320

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

321

■ W

E B

2 .0

A N

A L

Y T

IC S

Even with a bookmark to this website, getting there requires a two-step process (one to open the browser and one to access the bookmarks). The team does not play every day, so there may or may not be new content for me. With RSS, consuming con- tent about the Indian cricket team is significantly easier:

1. Go to the Cricket page on news.bbc.co.uk and click the orange RSS icon in the URL window, as shown in Figure 12.20.

Figure 12.20 BBC Sport: RSS sign-up for cricket in India

2. My feed reader (which happens to be the application Feedreader) automatically opens, asking me to add the XML feed to my subscriptions. See Figure 12.21.

Figure 12.21 BBC Sport: adding the feed for cricket in India

3. Click OK. You are finished.

Now rather than me checking for new stories about my sports team, my Fee- dreader will automatically check with the BBC Sport website every day. If there is a new story, it will show up in my Feedreader, and I never have to visit the website again. See Figure 12.22.

Figure 12.22 Feedreader view of the latest content

30652c12.qxd 5/1/07 12:48 PM Page 321

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

322

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Although the preceding example is of a sports fanatic, and most of the buzz around RSS is for blogs, you can imagine how this can apply to your company web- sites. If you have a content site, this is a no-brainer. You make it easier for customers to get to your content, and you will have more customers. But you can also think of other applications. For example, I have a D-Link router that is not the greatest on the planet. Rather than me proactively checking for the latest updates on the D-Link sup- port websites, it would be great if D-Link offered an RSS feed for the support page. Then, if there were new content, it would show up in my Feedreader and I could download the patch. Figure 12.23 shows the D-link support website and illustrates its lack of support for RSS.

Figure 12.23 D-Link support could provide RSS

The challenge with RSS is that your visitors never visit your website. They are not creating entries into your web log files and they are not executing your JavaScript tags. So your customers are consuming all the content from your website, but your web analytics application data is incomplete because it has no data for these customers.

It is not just that most of the web analytics applications are blind to this data, but also that there are no current standardized ways to track these metrics. You can sign up for RSS using feed readers (as I did in the preceding example) or you can sign up with major services such as Bloglines. My feed reader will say I am one reader, as will Bloglines. The issue is that hundreds of people use Bloglines, but they show up as only one reader in many statistics. Likewise there are many feed readers that each have their own way of behaving, making it hard to have standardized tracking (text readers, such as Feedreader shown in Figure 12.22, strip out all the HTML and present only text). This will obviously change as RSS becomes a much more dominant way of con- suming content.

Tuesday: Learn about Emerging “Standards” for RSS Tracking

RSS is still not very integrated into most websites and the analytics tools that are being used by the websites. There is no standard when it comes to tracking RSS. We have a couple of emerging options.

First, you can use standardized feed management services (for example, Feed- Burner). You can outsource your feed management to services such as FeedBurner, which can not only help you manage your feed subscriptions but also provide you with the ability to do various cool things with your feed. Such services will also report the core statistics for your RSS feed, as shown in Figure 12.24.

30652c12.qxd 5/1/07 12:48 PM Page 322

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

323

■ W

E B

2 .0

A N

A L

Y T

IC S

Figure 12.24 FeedBurner feed management service

Additionally, some vendors, such as Visual Sciences, are integrating feed report- ing into their standard web analytics application services. You can add a simple JavaScript tag, and as your feed runs around the world, the data is sent back to your Visual Sciences data collection services and included in your standard web analytics reporting. This will become more common in 2007 and beyond.

One of the challenges of this emerging medium is that cookies are not accepted by the feed readers. Some don’t accept JavaScript. There are no page views in the tradi- tional sense, and of course no sessions that we are so used to. All this means that it is hard to match existing web analytics metrics in an apple-to-apple comparison. Hence as you had learned in Chapter 7, “Month 2: Jump-Start Your Web Data Analysis,” we computed the metric Unique Blog Readers. It is a very different metric from any that are available in current web analytics tools.

You’ll spend the rest of the week learning about some of the metrics that are currently de facto standards. You can use these to measure success of your own RSS efforts.

N o t e : The difference between FeedBurner and Visual Sciences is that Visual Sciences will not manage

your feeds. For that you can use one of the free services online or implement your own on your website.

30652c12.qxd 5/1/07 12:48 PM Page 323

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

324

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Wednesday: Track Subscribers and Reach

As always, the first thing you want to know is, “Is anyone out there consuming my wonderful content?” Because the standard rules and metrics of the web analytics world don’t quite apply to the world of RSS, we will define and measure two new metrics: the loose equivalents of visitors and unique visitors.

Subscribers

For a given time period, this is the number of unique feed readers (online or desktop based) that have requested your feed. It is an approximate measure of the number of individuals who have opted to receive regular updates from your website, as shown in Figure 12.25.

Figure 12.25 Feed subscriber report from FeedBurner

The best way to think of this number is to imagine it as the subscriptions to a magazine or newspaper. This is simply the number of people who are getting your magazine or newspaper. As with magazines, the same person could have a subscription at work and at home, so a small amount of duplication is possible.

With FeedBurner it is also important to remember that this number includes your customers who have signed up to get the feed via email. In Figure 12.25, for my Occam’s Razor web analytics blog, it is 10 percent of the subscribers.

Track your subscribers over time. At least in this case, up and to the right is good, as shown in Figure 12.26.

30652c12.qxd 5/1/07 12:48 PM Page 324

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

325

■ W

E B

2 .0

A N

A L

Y T

IC S

Figure 12.26 Subscriber metric trend from FeedBurner

Reach

Reach is the total number of subscribers who have viewed the content in a feed during a given time frame. To continue our magazine subscription metaphor, this is the num- ber of people who have opened the newspaper and actually read a story (how wonder- ful is this—you can’t even track that with offline media!).

Figure 12.27 Reach metric trend from FeedBurner

The metric Reach is also a directional measure of engagement: of all the people who have signed up for your feed, how many are consuming the content and are hence engaged.

Thursday: Measure Reads and Clicks to Website (Referrals)

Now that you are comfortable with the metrics related to people/customer counts, you can move to the next step of understanding what specific piece of content that you are publishing is being consumed and is causing referrals to your website (where you hope- fully have an opportunity to sell products and services or simply engage your RSS sub- scribers in a deeper way).

N o t e : Some applications and practitioners use the term readers in place of reach. However, I prefer reach,

which is gaining some traction in the industry.

30652c12.qxd 5/1/07 12:48 PM Page 325

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

326

C H

A P

T E

R 1 2

: M

O N

T H

7 :

C O

M P

E T

IT IV

E I

N T

E L

L IG

E N

C E

A N

D W

E B

2 .0

A N

A L

Y T

IC S

■

Reads ( Views)

Reads (Views) computes the number of times each RSS item you have published is read by a subscriber during a given time frame.

In our magazine metaphor, this is a metric that would measure exactly which of the 90 stories in the magazines have been read by your subscriber. (I admit that in the cool world of Web 2.0, this is essentially measuring something equivalent to page view, so the page view is not really dead!)

This is a great way to understand which content is more popular with your sub- scribers and which is not. You can use this data to prioritize the areas that you can focus on to fuel your subscriber growth.

Clicks to Website

Clicks to Website is a metric that computes which RSS content is driving traffic to your website during a given time period.

In our magazine metaphor, think about being able to measure which scandalous high-profile celebrity story is causing your magazine subscriptions to soar, or more people to sign up for your sister publications.

A number of businesses publish RSS feeds of their content, but often they will have only partial content published in the hope that you’ll be tempted by the preview content and click over to the website, where you will promptly empty your wallet to get a paid subscription or make some other purchase (such a “non-social-media” thing to do!). There are other reasons why it might be good for your business to have your subscribers click over to the website. (For instance, subscribers might provide you with leads, sign up for newsletters, or consume non-RSS content on your site.)

Figure 12.28 shows a sample report that illustrates this metric. You can use Clicks to Website to understand customer interests and the kind of content that drives sub- scribers to your website. You can then optimize your articles or posts around that data.

Figure 12.28 Clicks to Website metric from FeedBurner

30652c12.qxd 5/1/07 12:48 PM Page 326

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

327

■ W

E B

2 .0

A N

A L

Y T

IC S

Friday: Analyze Subscriber Location and Feed Reader Type

Through your feed data, you also have an ability to report on two subscriber dimen- sions that could be useful for analyzing your visitors and improving your RSS feeds: location and feed reader type.

Location

You have an ability to report on the geographic locations of your subscribers, in case your business can optimize the website content or even their marketing campaigns based on this knowledge (see Figure 12.29).

Feed Reader Type

You also have an ability to report on the type of feed reader that your subscribers are using, as shown in Figure 12.29. This data is primarily helpful if you (or your IT folks) are interested in making special customizations to make your published content appear prettier in different feed readers (your subscribers will surely thank you for doing that).

Figure 12.29 Location and feed reader data for feed subscribers

Armed with this week’s information, you can give your RSS analytics a jump start. Wonderful reports and computed metrics that will impress your senior manage- ment team are within your reach. Remember that you can use free services such as FeedBurner to fill the gap that surely exists in your web analytics tool at the moment. Now you’re all set for at least a decent performance on your entry into the interesting world of Web 2.0.

30652c12.qxd 5/1/07 12:48 PM Page 327

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed

30652c12.qxd 5/1/07 12:48 PM Page 328

C op

yr ig

ht ©

2 00

7. J

oh n

W ile

y &

S on

s, In

co rp

or at

ed . A

ll rig

ht s

re se

rv ed