- 3,675,609 Users
- 11,593 Discussions
- 18,513 Comments
Database Cleanup: Improvements and Pitfalls of Cutting a Database in Half
It had been 3 years or more since we performed any database cleaning activities. And, many of our contacts were of questionable quality largely due to aggressively adding contacts from a list company as part of a concerted effort to increase visibility. Our contact count had more than doubled in the past 2 years from about 195,000 contacts to 430,000 contacts.
Marketing Challenge: Clean Up the Database
The main factor that made cleaning the database finally percolate to the top of my priority list was budget. It was contract renewal time and I urgently needed to cut from about 430,000 contacts down to 200,000. Doable, I thought, as I had more than 80,000 hard bounceback contacts and 30,000 unsubscribers. With 110,000 contacts quickly identified for deletion, I was about half way there but still needed to cut another 120,000.
I also wanted to clean the database to protect and possibly improve email deliverability. Though ReturnPath reported that my database was clean of spam traps, I wanted to confirm that. I also wanted to get rid of contacts who didn’t want my email. Doing so is a win-win-win; prospects aren’t annoyed, my brand image is protected, and my open rate improves. And open rate is increasingly monitored by email providers as a signal that your email is either legitimate or spam.
Solution: Segmentation and Re-engagement Campaign
In order to find the additional 120,000 contacts to delete, I created a segment to identify my least active 120,000 contacts. The segment contains a filter to sniff for any signs of life in a certain time period. Specifically, my filter pulls contacts in who have a “Date Created” value prior to 3 months ago and have been sent an email but NOT opened or clicked any emails, submitted any forms, or visited any landing pages, tagged pages, websites, or microsites within the last X months. I played with the number of months and landed on 9 months to hit my target of 120,000 contacts.
Most of what I had read on email database cleaning advocated eliminating contacts who are inactive 3 months, with some experts suggesting up to 6 months was ok. However, our specific b2b model suggests keeping inactive contacts much longer and so I opted for 9 months.
Now that I had identified the inactive contacts to remove from the database, I created and fed them into a re-engagement campaign to see if I could activate and save them. The campaign consists of 3 emails. The first advocates resources we have at our website to help the contact succeed. The Campaign Email Statistics operational report show that it was opened by 3,207 unique contacts. Part of the first email is reproduced below as viewed from the visual Click-through Report:
The second email features a call to action to update the contact’s preferences at our subscription center; it was opened by 4,346 unique contacts. The third and final email features a call to action to remain subscribed. If the contact clicks the link to remain subscribed, a blind form is submitted that subscribes the contact. Email 3 was opened by 1,788 unique contacts.
The campaign was configured such that if the contact did NOT open any of the 3 emails, they were added to a shared list called “Inactives to Remove from Eloqua”. See a portion of the campaign below:
Of 118,641 contacts that entered the campaign, 9,341 unique contacts (8%) opened one of the three emails and were saved from deletion. 109,300 contacts did not open any of the 3 emails and thus were added to the deletion list. Next, I deleted the contacts in the inactives shared list from the database. In this way, the number of contacts in the database was reduced from about 430,000 to 210,700, which was close enough for my purposes.
To analyze the results, I needed an apples-to-apples comparison. I settled on my web seminar email that has a subject line that didn’t change between the comparison periods and that I send each month to announce the web seminars for the coming month. I utilized the Email Analysis Overview report to come up with a set of metric averages before and after the cleanup. See the results below:
For the most part, the results look terrific. My total open rate increased 109.4% and my unique open rate increased 70.5%. Success! Right? Not entirely. Can you tell which metric bothers me the most?
Pitfalls of Cleaning Too Aggressively
The decline in the number of unique opens by 6.6% bothers me. If I hadn’t cut inactive contacts so aggressively, the results suggest I’d have had another 362 unique opens (5,507 – 5,145). That’s significant to me. Did I cut too aggressively? Perhaps. But I have to weigh the decline in unique opens against the gains in other metrics like open rates which protect and possibly improve deliverability.
If you have the budget, I recommend keeping your inactive contacts, except those that have never opened an email as those might be spam traps. Your contacts probably cost you a lot to acquire and it seems economical to add contacts to your contract limit, especially when adding contacts within your current plan level (i.e., Basic, Standard, Enterprise).
If you can keep your inactive contacts, I’d recommend a hybrid email approach. Create an inactive filter and apply that to most of your email campaigns so that you don’t email inactive contacts on a regular basis. Just email them once in a while to see if their needs have changed such that your marketing messages now hit their radar. By taking this approach, you improve your overall open rate and other activity metrics while keeping inactives in the loop from time to time.
Helpful Eloqua Courses
The tactics I employed were learned in the following Eloqua courses:
- Best Practices: Email Deliverability and Privacy
- Insight for Reporters
- Fundamentals of Segmentation
- Advanced Segmentation
- Blind Form Submits
- Fundamentals of the Campaign Canvas
- Fundamentals of Emails