Hack45.Measure Affiliate Marketing


Hack 45. Measure Affiliate Marketing

Many online marketers take advantage of affiliates and affiliate programs to drive traffic to their sites without considering everything that can be measured.

Affiliate marketing is a very powerful customer acquisition strategy, essentially paying people a commission to evangelize your products or services and then sending traffic your way in hope that they'll complete the transaction. Your web measurement application should allow you to track affiliate campaigns, much like a banner ad or email message. Tracking your affiliate campaigns in as much depth as possible helps you learn from the creativity of others as you plan your own future marketing activities.

3.10.1. Pay for Clicks, Leads, or Revenue: Which Is Better?

Most affiliate management programs (Commission Junction, Linkshare, and Performics) require that you pay on either a per-click, per-lead, or commission basis (Table 3-7).

When you're paying for clicks, you might be paying for traffic that never actually visits your web site [Hack #51], or that never actually converts into a customer [Hack #39]. The best example of this type of affiliate relationship is the Google AdSense program. On a per-click basis, affiliate publishers are paid by the amount of traffic they generate. The more clicks generated, the more the affiliate gets paid. In this model, you want to be sure you're watching your affiliate conversion rate and the amount of revenue individual affiliates generate.

Where you're paying for leads, you want to be careful to watch out for poorly qualified leadsvisitors who are unlikely to convert into revenue or other value for your business. If your affiliates send you poorly qualified traffic, the cost of all generated leads will be greater than the amount of revenue generated by those customers. In this model, you want to be sure you are tracking your profit or loss per lead to make sure this traffic maintains profitability.

The ideal financial relationship between your site and your affiliate network is pay-for-performance, for which you pay a commission per sale generated to your affiliates. This strategy can be measured only through value-based metrics, captured at the point of sale and tied back to the originating affiliate. With this type of relationship, there is no need for a loss column on any reports because the only payments that will be made will be to affiliates that generate revenue.

Table 3-7. Comparison of each type of affiliate marketing payment program

Type of program

Advantages

Disadvantages

Pay per click

Easiest to understand and monitor

Many clicks never make it to the site, and quality of clicks can vary greatly, driving up per-click and per-order costs.

Pay per lead

Guarantees that you gain more information about the prospect

The leads you get may not be "good" leads, thus driving up the cost of real leads.

Pay per performance

Guarantees that money goes in your pocket before any payment is made to the affiliate, and helps foster a mutually beneficial relationship

Usually the most expensive on a percentage of revenue basis, and sometimes difficult to create cost effective partnerships.


Regardless of what you work out, there are a handful of standard measurements you should take to determine the success of your affiliate programs.

3.10.2. Use Web Measurement to Identify Affiliate Successes

Despite the fact that many affiliate programs come with some type of measurement tool built in, you owe it to yourself to validate those numbers using your web measurement application. Additionally, affiliate programs usually measure only what they pay on (clicks, acquisition, revenue), but the key insights you hope to gain about affiliate marketing will come from a deeper level of measurement, the kind nearly any web measurement tool provides.

The following are some of the key measurements you should be making on your affiliate traffic activity.

3.10.2.1 Paying for clicks? Determine your per-affiliate click-through rate.

Your web measurement application should be able to quickly summarize the number of visits each affiliate sends your way, based on a unique URL string or query string parameters. For example, clicking a link on your affiliate's web site might bring direct visitors to http://www.mysite.com/landingpage.htm?source=affiliate1. The presence of thesource=affiliate1 would allow you to identify that the visitors came fromaffiliate1and track the visitors through to conversion.

While the number of respondents is often a poor proxy for clicks unless your affiliate management application provides you the necessary data, respondents and a visit-based measurement will have to do. Make sure you build a click-through rate using impression and click data reported from your affiliates or estimated impressions and respondent data. Even if someone else is managing the measurement and payments based on clicks, taking these measurements yourself will help you identify fraud and other activities you don't want to pay for.

3.10.2.2 Paying for leads? Measure lead generation rates.

If you are paying your affiliates on a per-lead basis, you'll need to track when a form is successfully completed. The easiest way to do this is to identify the "thank you for your submission" page so that you see the original referrer of the traffic and the number of times that referrer produced traffic that landed on that page. Most web measurement applications allow you to set a goal page for analysis back to the referrer (Figure 3-9).

Figure 3-9. Visitor fall-off from affiliate to conversion


An even better strategy would be to create a conversion funnel [Hack #59] that lets you see on a per-affiliate basis how many visitors start and finish the lead generation process. Especially when yours is a more complicated process, identifying where abandonment occurs can help you improve the quality of your relationship with your affiliates. Remember, the best affiliates want to hear your suggestions for how to improve their message, since it increases their revenue stream at the same time!

3.10.2.3 Paying for performance? Paying commissions based on sales.

Ideally, you are paying your affiliates a percentage of each sale they generate. This can vary depending on your model, but it is the recommended model for all affiliate relationships. This requires that your web measurement application be able to record the specific dollar amount that was given to you by the customer, the rate at which the affiliate earns their commission per sale, and the ability to tie the sale back to the original referrer. Table 3-8 shows an example of this information.

Plan on generating a report for "conversion and revenue by affiliate" for yourselfeven if the affiliate management program you use generates onefor validation. If you're really sophisticated and are able to input the percentage you pay each affiliate, you're likely to be able to get any moderately sophisticated web measurement application to calculate your total affiliate sales and payments for you as well.

As you can see from Table 3-8, if you were paying per lead for the "Free Guide to Web Analytics," you would be losing money. By measuring all levels, you can determine the optimal structure for each affiliate relationship.

Table 3-8. Sample affiliate analysis showing the relationship between affiliate performance and profit

Affiliate name

Rate

Responses

Leads

Conversions

Revenue

Cost

Profit

Web Measurement Blog

25%

435

400

10

$500

$125

$375

Analytics Bible

5%

34

14

0

$0

$0

$0

Librioteca de Web Analytics

10%

3,423

1,245

17

$849

$85

$764

Web Analytics Org

25%

3,404

1,789

56

$2,797

$699

$2,098

Association of Analytics Geeks

12.5%

4,543

58

58

$2,897

$362

$2,535

Free Guide to Web Analytics

15%

9,845

8,765

0

$0

$0

$0

Metrics Matter

20%

4,932

2,805

132

$6,593

$1,319

$5,275

TOTALS

 

26,616

15,076

273

$13,636

$2,590

$11,046


3.10.2.4 Other useful metrics used to measure affiliate marketing.

In addition to the key metrics we describe above, you should be thinking about your affiliate traffic like any other valuable referring source and taking advantage of it. If you have robust visitor segmentation tools [Hack #48] in your measurement application, create a segment from affiliate visitors and see how their activity differs from other acquisition segments. Make sure you're looking at KPIs relevant to affiliate visitorsaverage time spent on site, average page views per visit, percentage new versus returning visitors, percent visits less than 90 seconds, and percent interested and committed visits (for a more complete list, consult Chapter 7)as well as important measurements and lists like the top 25 pages viewed and top entry and exit pages.

3.10.3. Tying It All Together

Make sure you treat your affiliate partners like the valuable revenue stream they can be and the visitors they drive like the potential customers they are. Just because affiliate marketing is made easy by aggregators such as Commission Junction, Linkshare, and Performics doesn't mean you should take your affiliates for granted.

Finding beneficial affiliate relationships can be difficult to do if you aren't willing to meet halfway between their wants and your needs. You may need to try out relationships for a month to determine what the affiliates' expected revenue would be over the entire year. Don't be afraid to run a test rollout and then renegotiate the contract based on the metrics you collect. Let the affiliate know your sales expectations and their progress throughout the relationship. Constant communication about the resulting metrics can only benefit both parties.

Also, if possible, have the affiliate track their progress using their own measurement solution and compare notes about results. There will be differences between measurement systems, but this will help generate trust, and that is something that can't always be measured with a hack.

Dylan Lewis and Eric T. Peterson



    Web Site Measurement Hacks
    Web Site Measurement Hacks: Tips & Tools to Help Optimize Your Online Business
    ISBN: 0596009887
    EAN: 2147483647
    Year: 2005
    Pages: 157

    flylib.com © 2008-2017.
    If you may any questions please contact us: flylib@qtcs.net