Web Metrics

MARKETING TRACK – Step 3

Topic: Web Metrics Pt.1 — We test 26 different web metrics tools to determine the simplest, most accurate way to capture the numbers you need.

Test Number: #030202-WA

Word Count: 6300+

Focus:

SECTION 1 (Five Questions)

1. Can you trust the click reports provided by Google and Overture?
2. Why don’t the reports from the pay-per-click engines match the reports from our analytics programs?
3. Can you trust the numbers your web analytics program is providing?
4. Why is it so difficult to get accurate metrics from Yahoo! Store?
5. How do you cut through the confusion and find a web analytics tool that really works?

SECTION 2 (Outline)

1. What do you (really) need to measure and analyze? (Five questions you need to answer)
   1. Thirty nine activities you can measure
   2. Eleven activities you probably should measure
2. How do you select the best metrics tool?
3. What are the available metrics tools?
   1. Five simple (homespun) solutions
   2. Eight ASP solutions
   3. Twelve software solutions
   4. Lab-based solutions
4. What metrics tools do we use at MEC? (Eight ways we get the numbers we need)

Credits:

1. Writer – Flint McGlaughlin
2. HTML Designer – Cliff Rainer
3. Contributors – Brian Alt
   Aaron Rosenthal

Study the Data from this test

We have three (aggravating) questions:

(1) How could three separate web metrics programs be given access to the exact same numbers (526 orders), and yet deliver a set of metrics that are 276 percent apart? (2) How can Google’s Adword Select program report 7,706 more clicks than our actual log files report? (3) More importantly, how can a marketing team make effective decisions if they don’t have accurate numbers?

Over the last six months, researchers at MarketingExperiments.Com have tracked more than 36,000 orders, studied 800 pages of advice, analyzed more than 10 GB of data, and reviewed 22 web metrics products.

We were searching for simplicity.

As Judge Oliver Wendell Holmes remarked, “I would not give a fig for the simplicity this side of complexity, but I would give my life for the simplicity on the other side of complexity.”

Here at the lab, we have grown painfully aware of the need for a simple approach to web metrics.

If we have discovered anything significant, it is probably this: effective, serious-minded merchants, across the Net, in diverse industries, are struggling with inadequate data.

Their (understandable) ignorance is costing revenue and growth.

The Internet was supposed to provide a wealth of useful data, but for many, it is a quagmire of disconnected, inaccurate numbers.

Is there a way to cut through the web metrics confusion?

Here is what we discovered.

Can you trust the click reports provided by Google and Overture?

Most marketers that we have interviewed experience vague misgivings about the amount of money they are spending on the pay-per-click engines. Are they being overcharged?

We tested 50,000 plus clicks from the two leading PPC engines. Here is what we discovered:

What You Need To UNDERSTAND: There is a reporting variation of 19 percent! That is to say, if your cost per click was $0.49, you could have been overcharged by $3775.94. (*1)

What You Need To UNDERSTAND: There is a reporting variation of 8 percent. That is to say, if your cost per click was $0.49, you could have been overcharged by $200.41. (To get charts like this for your website, click here: http://216.110.184.95/cgi-bin/pl/pl.cgi?syndicate )

We are not suggesting that these pay-per-click engines are engaged in dishonest reporting. But an error can cost you as much money as a lie.

And in times past, we have seen evidence of significant errors on the parts of at least one PPC engine.

The moral of the story?

Get your own (reliable) metrics in place.

Why don’t the reports from the pay-per-click engines match the reports from our analytics programs?

There are at least two reasons:

1. If someone (1) clicks on your link in Google Adwords, (2) visits your site, (3) returns to Google, and then (4) hits the back button, Google counts this as another billable click.
2. Both Google and Overture have many small affiliates. Without a special tracking code, a log file analytics program such as ClickTracks cannot discern whether a visitor came from the main search engine or its affiliate.

NOTE: You can view an updated list of Google affiliates in our Research Log at:

http://mecgroups.com/mec_research_log/archives/000029.html

Can you trust the numbers your web analytics program is providing?

We’ve established that tracking your clicks can be difficult, but what about tracking your orders?

In our next experiment, we studied 526 orders to determine how many of them originated from a Google Adwords pay-per-click campaign.

This campaign was for some 1000+ products offered in a Yahoo! store. We measured the results with two programs:

1. Yahoo! Management Interface – The web stats program built into the Yahoo! Store platform.
2. HyperTracker Pro Version – A hosted web analytic solution specifically designed for measuring click-based campaigns (we discuss this program in more detail in SECTION 2).

We compared the numbers from both programs with a manual (oh-so-tiresome) URL trace of every single order. The results were frightening:

What You Need To UNDERSTAND: Yahoo!’s reporting was “off” by 276 percent.

How much money did we (really) make or lose with this Google campaign?

We ran a simple ROI report based on the three different sets of numbers. One thing was clear: if the merchant had relied on Yahoo! or HyperTracker, his analysis could have distorted future campaign decisions:

Any merchant relying on the Yahoo! numbers would likely panic and terminate his Google campaign. But in reality, this campaign showed economic promise. Here are the (mostly) accurate numbers: (*2)

What You Need To UNDERSTAND: According to Yahoo!, the merchant was losing 52.19 percent. But in actuality, the campaign generated a return of 31.89 percent. (To get charts like this for your website, click here: http://216.110.184.95/cgi-bin/pl/pl.cgi?syndicate)

Armed (at last) with accurate numbers, our research partner grappled with the next vital question:

Should he continue with his Google campaign?

The answer is not as simple as it may seem. The decision would be based upon a single vital factor: his LIQUIDTY.

At MEC, we have a special metric. We call it the Liquidity Index. It is a single number that answers THE cash flow question:

Will I recover my campaign expenses within thirty days?

To be sure, a marketing investment can be sound even if the money is recovered over a 12-month cycle. But in today’s ferocious economy, many marketers cannot afford to wait for 12 months.

Indeed, many marketers don’t even have a fixed budget. They hope to recoup their direct costs every thirty days.

Our research partner had two obvious options:

1. IF HIS BUDGET WAS GENEROUS – Buy all available traffic of this quality. If your margin is 40 percent, $39.49 is not too much to pay for a customer that will spend $130.20 in the first year. And every customer acquired is a customer the competition doesn’t acquire.
2. IF HIS BUDGET WAS LIMITED – “Freeze” the campaign to avoid receiving a end-of-month bill that he was unable to pay.

We recommended a third option: (1) Refine the Google campaign, (2) eliminate unprofitable keywords, (3) enter negative keywords, (4) improve the landing page, and (5) monitor the results closely.

This third option has proven to be the best choice.

Unfortunately, many marketers don’t have enough accurate information to discern a “third option.” Nor do they have the confidence to buy all of the available traffic for a campaign.

While preparing this report, we spoke to one merchant who admitted, “I’ve never known how much my Google campaign was really yielding, so I just low-balled the bids.”

This merchant Grossed 1.7 million last year with his homegrown Internet operation… but how much market share did he lose by underbidding on profitable search terms?

Why is it so difficult to get accurate metrics from Yahoo! Store?

Yahoo! Store offers one of the best hosted e-commerce platforms on the Internet, but their site metrics are sorely lacking. There are at least seven problems:

1. Yahoo! does not grant access to its log files.
2. The referral link section, entitled “References,” is absolutely unreliable.
3. Yahoo! is an affiliate of Overture, and you cannot determine if your orders are coming from the Yahoo! network or from a separate Overture campaign.

   NOTE: Our research partners report that they are (sometimes) charged a Yahoo! Shopping Commission for orders that came through their pay-per-click campaigns.

4. Yahoo! will not let you place the necessary code on the order confirmation page to set up a simple tracking system.
5. There is no easy way to view, download, and analyze a complete list of your Yahoo! store orders.
6. Yahoo! often displays orders with no referring link and no entry point.
7. Yahoo!’s technical support is outsourced and unable to assist you with important questions about your metrics.

We are working with one of our research partners, Don Cole of ystoretools.com, to develop an effective analytics program for Yahoo!

How do you cut through the confusion and find a web analytics tool that really works?

The Scottish writer, Andrew Lang, once commented, “He uses statistics as a drunken man uses a lamppost — for support rather than illumination.”

The goal of web analytics should be illumination. And…

Knowing what we DON’T need to measure is at least as important as knowing what we DO need to measure.

The key is focus.

Effectiveness is not the result of measuring the numbers right; it is the result of measuring the right numbers.

Your company’s time and energy is at a premium. You need to get the “most” for the “least.” This rigorous focus not only concentrates your resources; it also clarifies your objectives.

When you cut through all the confusion, there are really only four elements you can measure:

1. The AMOUNT of activity on your site – page views, visitor sessions, returning visitors, etc.
2. The SOURCE of that activity – referrers, search terms, languages, countries, organizations, etc.
3. The NATURE of that activity – entry pages, exit pages, browsers, platforms, JavaScript versions, cookie support, screen resolutions, page refreshes, page load errors, average time per page, etc.
4. The RESULTS of that activity – click trails, most requested pages, number of page views, signups, orders, etc.

In each of these four categories, there are only a few numbers that will truly matter.

Once you have collected the numbers, it is time for the second step: analysis.

Lest this ominous word frighten you, most of your analysis can be painlessly accomplished with nothing more than a one-page spreadsheet.

Section 2 (Continue…)