Chapter Six

Testing and Tracking

By Ben Hart

I often chuckle when people ask me if I think their idea will work.

My answer is always the same: “I don’t know.  Test it.”

People absolutely hate that answer. But it’s the truth.  The only way you can know if something will sell is to try to sell it, and to try different ways to sell it.

The most important word in marketing is “test.”

You must constantly test ads, headlines, offers, landing pages and products.  This is a never-ending process.

I’ve been in direct marketing for 20 years. I have mailed about a billion direct mail letters.

And I continue to be surprised by the results of tests.

Google AdWords allows you to conduct tests at lightening speed because the feedback is just about instant.  Within a few hours, you can know the results of your test.

I am refining and adjusting my AdWords campaigns all the time based on test results.

So how do you conduct tests?

Well, you could do it by hand with a calculator, a pencil and a yellow legal pad. 

But that would be enormously time-consuming.  And even then, you won’t be able to really get all the information you need.

Remember, information equals power.

The better and more exact your information, the richer you’ll get.  Having the right information is just like having the combination to the vault.

It’s super easy to get the money if you just have the right combination.  It’s very difficult if you don’t know the combination.

That’s what testing gives you – the combination to the vault.  The purpose of testing is to gather information that allows you to get rich by making all the right decisions.

The good news is Google gives you lots of tools to make it very easy for you to test and track the results.  Google even puts the results of your tests on spreadsheets for you.

Google has tools that allow you to track not just the performance of your ads, but also your keywords – and not just clicks on your ads, but also how they “convert” on your landing page and move through your sales process.

It’s all automatic.

You just need to spend a little time learning how these tools work.  But the little amount of time you invest here will save you hundreds of hours – and will allow you to gather key information that you never could have done by hand with a calculator and legal pad.

The Split-Test

The most basic test is a ­ split-test of your ads.

If you turn this option on, Google will rotate the two ads you are testing evenly throughout the day.  Make sure Budget Optimizer is turned off (which it should be if you’ve followed my advice). You will then see the results of clicks on your ad on a Google report, broken down by keyword and phrase.

You might see that one ad is the winner with some keywords, but not with other keywords. 

When you are conducting a split-test between ads, you must test one change at a time.

THE REASON: You can’t know what factor produced the winner if there are many differences between two ads.

The most important test for a Google ad is the headline.

That’s because your headline is the most important element in a Google ad.  Your headline is what your searcher notices first.  So that’s your most important test.

If you have more than one product to offer (and you should), run a head-to-head test between two products. Maybe more golfers want drivers than putters.

Test a word.

Does the word “free” in the headline boost or depress results.

Almost always the word “free” boosts the number of clicks on your ad.  But in some instances I have seen the word “free” depresses the number of clicks.

If you need a plumber now, “free” might depress results.  No one believes a plumber is free.  Or maybe the word “free” is attracting the wrong clicks – people who want your free offer, but won’t buy anything.

You must always test.  Never ASSUME anything.

So that’s the most basic test you can do -- a head-to-head test of ads that Google will rotate for you evenly throughout the day.

But that will only tell you how an ad is performing in terms of getting clicks.  This is called your “click-through rate.”

Internet marketers are intensely interested in click-through-rates.  No clicks means no traffic for your site. No traffic equals no sales.  So you certainly want clicks and lots of them.  But you want the right clicks.  Clicks from non-buyers just cost you money.

So the next element to track is your “conversion rate.”

Your conversion rate is what happens on your landing page.

I define conversion rate as my “Most Wanted Response.”  It the primary action you want your visitor to take when they reach your landing page.

In my case, it’s filling out my sign-up form to get my “ethical bribe,” which is the free ebook or “white paper.”

When someone fills out my sign-up form, that’s a conversion.

For another business, a conversion might be a sale if the landing page is a sales letter.

Clicks that don’t lead to a high-rate of conversion are worthless.  Actually, clicks that don’t lead to conversions are worse than worthless because clicks equal money you have spent.

So you need to covert.  You need to either get a lead or a sale. The end result of all your work is to generate sales.  So clicks and leads are no good if they don’t produce enough sales at the end of the marketing process.

You must track how your marketing process is doing at each key point.

Thank heavens Google has a tool that lets you do this automatically and almost effortlessly.  It’s called “Conversion Tracker.”

Google will give you some code that you paste on your “thank you” page.  The “thank you” page is the page that pops up as soon as your visitor fills out your sign-up form or makes a purchase.

Each time this “thank you” page or “order confirmation” page is opened, Google tallies the number for you and puts the numbers in a nice spreadsheet for you.

Now, this is the imperfect part.

Google gives you no automatic way to track the results of landing pages simultaneously.

That’s because an ad can only be linked to just one landing page at a time.

There are two ways you can overcome this problem.

You can let your ad run for a while. Once you collect enough data, you change the landing page, and then collect conversion data on that.

This will give you a fairly accurate answer.

As a general rule, I don’t like to conduct tests at different times because different times will yield different results.

A head-to-head test in direct mail means two different mail packages mailed at the exact same time.  If one package mails on Tuesday and another on Friday, this can skew the test.

Another method is to use Google’s ad split-testing tool.  You then create two almost identical ads – maybe one inconsequential word changed in the body text under the headline – a word no one would notice (so that Google does not think they are identical ads and flag them).

Then just connect each ad to a different landing page.  This will give you a good landing page test.

So that’s how you conduct tests for your Google Ads and landing pages.

How much data do you need to know the results of a test?

My rule is that I need at least 40 actions on each ad – whether clicks on the Google ad or sign-ups on my landing page.  In most cases, that’s enough data to determine a winner.  If results are not conclusive, you just keep letting the test run.

The more data you have, the more accurate your test.

If you flip a coin, how many flips do you need to achieve a statistical 50/50 ratio?

Certainly more than 10. I’ve seen flips of a coin come up tails 5 times in a row, even 10 times in a row.  If you flip a coin 50 times,  you will get closer to the 50/50 ratio. 100 flips will get you closer still. 1,000 flips and you will have a near-exact 50-50 ratio between heads and tails.

That’s how casinos make their money.  They know the exact odds of the games they’ve created.

That’s what polling is.  If you test a big enough sample of people, you will have a good measure of what public opinion is on an issue.  The bigger the sample, the more accurate the poll.

So that’s what you are doing with your tests.

You are conducting a poll of your Google ads and landing pages.  Your clicks and your conversions are your answers to the poll you are conducting.

Ad testing and polling are really the same science.  But a test of your Google ad and your landing page is much more accurate than a poll you read about in the newspaper.

Why?

Because an opinion poll is just asking people their opinion – let’s say about political candidates many months before an election.  How they actually vote on Election Day can be an entirely different story.

But with this test of your Google ad and your landing page, you are measuring specific actions that people are actually taking.  It’s the difference between asking people what they think of your product and asking them to buy your product.

This is why testing is so key to your success.

If your test results show success, your marketing campaign absolutely will be successful – no doubt about it.  It’s as certain as knowing the combination to the vault.

What’s truly awesome about Google’s tracking technology is that it not only tracks your ads’ performance and the conversion rate on your landing pages, it also tracks the performance of each of your keywords.

What you will find is that some keywords and phrases deliver lots of clicks, but then don’t covert on the landing page.

Those who are not converting are just costing you money with their clicks.  Very likely you will want to weed out certain keywords that generate lots of clicks, but few conversions.

And Google gives you all this data, by ad campaign and keyword in a nice spreadsheet that you can read and absorb at a glance.

Sure beats the heck out of tracking it all by hand!

You don’t want to be spending your time entering data and creating spreadsheets.  Let Google do that.  You want to focus on running your business.

But don’t make the mistake many make and put Google’s conversion tracking code on your landing page. Or your report will show a 100% conversion rate.  Paste it on your “thank you” page or “order confirmation” page.

If you have several levels of conversion or actions you are tracking, you will want to paste Google code on each page that comes up after your visitor has performed the action you want.  So, for example, you would paste tracking code on both the “thank you” page that comes up after they sign up for your free offer and again on the “order confirmation” page that comes up after a purchase.

Google will then give you reports on how your ads and keywords are doing – both in terms of generating leads and generating sales.

The Power of Google Analytics

Far more powerful than Google’s Conversion Tracker Tool is Google’s newer Analytics Tool.

Analytics does everything Conversion Tracker does, plus a whole lot more.  Analytics can tack the keyword source and pay-per-click ad source of your visitors just like Conversion Tracker.  But Analytics will also track visitors from all your other advertising sources.  Analytics will even track your traffic coming from your organic traffic and your offline advertising.  If you have PPC ads on Yahoo, Analytics will also track those for you.

Analytics will measure all levels of conversion you are interested in.  Analytics will also tell you how long visitors are spending on each page of your website and which page seems to be triggering exits from you site.

Analytics will also track the productivity of your traffic by geographic location.  You will find that traffic from some regions are much more likely to buy than traffic from other regions.  For example, your traffic from a wealthy area such as Scarsdale will be more likely to buy than traffic from Newark.

You need to know not just what sources (i.e. your ads, keywords, and organic search) are bringing traffic to your site, you need to know where your buyers are coming from. 

Google Analytics will slice and dice information all kinds of ways and give you a myriad of reports on exactly what your traffic is doing on your website, broken down by keyword source, ad source and geography.

The information on how your site is performing Analytics will give you is extraordinary, and very useful as you fine-tune your website and selling process.

Just like with Conversion Tracker, Analytics is very easy to set up.  You just copy and paste Analytics code on each page of your site you want to track.  If you know nothing about HTML and how to do that, just ask your web guy to do it. 

If you want to track just simple conversions, go with Conversion Tracker.  If you want much more elaborate reporting on what’s happening on your website, use Analytics.

With both tools, Google puts a cookie in the computer of your visitor that lasts up to 30 days.

So if someone signs up to get your free offer, and then come back to your site five or ten days later to buy something, both Conversion Tracker and Analytics can track that and spit out great reports for you.

What you will learn from both Conversion Tracker and Analytics is where the problems (and successes) are in your selling process.  If you are getting a lot of people signing up for your free offer, but you are closing few sales, then you know you have a breakdown.  Either your sales presentation is no good, your offer is not compelling, or people don’t want what you are selling.  Or maybe there’s something wrong with your order form.

Google Conversion Tracker and Analytics will show you exactly where people are bailing out of your marketing funnel. 

Remember, information is power.  The first step to fixing a problem is know where and what the problem is with your conversion and selling process.

You can use Google Analytics even if you are not an AdWords customer.  You can also have Analytics and Conversion Tracker running at the same time.  The two methods of tracking activity on your site do not interfere with each other.

I like to run both because the reports are different.

I don’t always need the Ferrari-level tracking that Analytics gives me.  You could spend all day studying data available in Analytics. Often I just want basic tracking information that Conversion Tracker gives me.

When you do your tracking and testing of your ads, be sure you are measuring the right data.

Don’t make this more complicated than it needs to be.  Some people go overboard and track too much data.  You will want to know . . .

1) How much it’s costing per click.

2) How much it’s costing to get a name and email address -- a lead.

3) And how many names (leads) you need to generate a sale, which gives you your average cost to generate a sales from a first-time buyer.

4) Average Long-Term Value (LTV) of finding a fisrt-time buyer.

That’s how to track your return on investment (ROI).

You should also test and track headlines (both on your ads and landing pages) and what keywords and phrases are the most productive in terms of leading to sales.

And that’s about it.  But I can’t emphasize the importance of testing enough.

Everything we know about what works and doesn’t work in direct marketing is the product of trial-and-error -- of testing.

And much of what we know about what works and what does not work goes against what we would have predicted.

For example, you might think short sales letters would work better than long sales letters in direct mail.

But testing shows that long sales letters work better than short letters 80 percent of the time -- or more -- particularly in prospecting.

Long letters on sales websites work better on the internet also, most of the time.

The only way we know this is true is because of testing.

Never take anything for granted in direct marketing.  Never assume anything.  Make educated guesses.  But test everything.  Test all your assumptions.  Two different headlines can produce a 1,000% difference in results.  And often the headline you think will work best is the one that pulls one-tenth what the other headline pulls.

Which of these two headlines do you think generated the highest “click-through-rate”? 

Simple Diet Plan, Free

 

Easy Diet Plan, Free

The headline with the word “simple” generated 72% more clicks than the headline with the word “easy.”

That’s what a change of just one word in a headline can do – even when the meaning of the word is almost the same.  You can never do enough testing.  The more testing you do, the more profitable your marketing will be.

But I also don’t want to overwhelm you with information here -- or make marketing with Google AdWords sound more difficult than it really is.

Don’t make the perfect the enemy of the good.

What I urge you to do is to just get on Google AdWords and start playing with it.

If you don’t have your tracking and test matrix’s set up perfectly, don’t worry too much.  Just start trying things. 

You’ll be able to tell pretty quickly what’s working and what’s not working, really just by eye-balling your clicks and conversions.