Tuesday, February 2, 2010

Evaluating Marketing Automation - Data Management


Continuing on a theme that received great feedback, I wanted to provide another real, down in the details, way to evaluate the various claims in the marketing automation field. Last time we looked a way to ensure that a provider could have the performance needed for your marketing goals - a quick and simple upload that will test actual marketing automation system performance.

In this post, it's worth taking it one step further. Getting marketing data into a platform is one thing, but if the data is messy (and what marketing data isn't), it will not be of much use. If, for example, your marketing database has 100,000 names in it, and the titles are just as they were written, such as:

  • VP Marketing
  • V.P. Mktg
  • Vice Pres Marketing
  • Marketing Vice President
  • Mktg VP

and you are asked to build a list of Vice Presidents of Marketing to target, how many will you find? 300? 800? We've seen many situations where dirty data returned 300 names, but the same query against cleansed data returned 17,000 names. Proper management of data makes a huge difference in your marketing results.

So, how do you test for this when considering a marketing automation software investment?

Quite simply - ask, in a demo, for each vendor you are considering to run a quick test. Here is a sample CSV file with typical marketing data. Titles, states, and countries are as they would be in a normal marketing or CRM database. The data is kept simple, and the titles are mostly in sales, marketing, and finance, while the addresses are in Canada, US, and UK.

Have each vendor run the following test for you:
  • Upload the sample file
  • Clean up the country fields so that US, USA, U.S.A, as well as the variations of Canada, and England/UK are normalized
  • Clean up the "raw" job title fields to two new fields for "level" (VP, Director, etc), and "role" (marketing, finance, etc) so you can properly segment
  • As a bonus, see if they can correct the missing leading "0" on New England zip codes - removed by Excel in many marketers' data files
When it is uploaded and cleansed check the data to see the following:

  • The only countries in the file are "USA", "GBR" and "CAN" or however you chose to normalize the country data
  • The people can easily be filtered by role into "Marketing", "Sales", or "Finance"
  • The people can easily be filtered by level into "SVP", "VP", "Director", or "Manager"

Many marketing challenges come from bad data. An inability to do proper segmentation, personalization, lead scoring, or analytics can quickly result if you are not able to standardize and normalize the data in your marketing database. To avoid getting into this situation, it's worth having the marketing automation provider you are thinking of choosing run through this quick test with real sample data.
BOOK
Many of the topics on this blog are discussed in more detail in my book Digital Body Language
SOFTWARE
In my day job, I am with Eloqua, the marketing automation software used by the worlds best marketers
EVENTS
Come talk with me or one of my colleagues at a live event, or join in on a webinar

4 comments:

David Raab said...

Tremendously helpful Steve. Thanks.

B2B Data Marketing UK said...

Some good points expressed here =)

marketing automation solutions said...

Very nice post and sample data to test out systems. A few comments here, I think it's rare to find data normalization offered as a feature in a marketing automation system. Do you feel it's best to have this feature here or have it inherent within the CRM system, such as Salesforce.com? To me, this seems like more of a CRM feature.

Steven Woods said...

I'd actually say that it *should* be in marketing automation. Generally, marketing gets the messiest data - web forms, lists from shows, etc, and should ideally just present clean data to the CRM system.

That being said, you're right that many marketing automation systems are a bit weak in this area.