I wrote this blog about direct mail almost a year ago. I’ve got a few updates:
Updates:
- Population Re-weighing. You might want to constantly re-weigh your mailable universe against the entire mailable universe. Otherwise your model or recalibrated model might become myopic and start chasing it’s own tail and spin out of control. The telltale signs are that your response rate is dropping but your conversion rate is slightly higher.
- Data is everything. Work with a credit bureau that can provide you with the best sets of variables especially trending variables that will always give you a leg up. It is even better If a particular bureau let’s you download the non-PII data to exercise your response model.
- High Utilization % are not revolvers. Some of our clients are targeting high utilization % consumers thinking that they are revolvers, it’s not always the case. Work with your analytics team or credit bureau to help you identify true revolvers.
Original Blog:
- Volume. To obtain meaningful results, you have to commit to a meaningful volume. A few thousand pieces here or there will lead to upsetting results and might prematurely end your efforts in using direct mail as a acquisition channel.
- Avoid Using Only Marketing or Demographics Data. They don’t work and don’t let cost lure you in. Always use a credit bureau and actual credit attributes.
- Use Multiple Credit Bureaus. the coverage between the 3 prime bureaus are not the same. Even if the same prospect can be found across all bureaus, the depth of their credit profile are quite different.
- Response Modeling Is Not Enough. If you are fortunate to have a team of data scientists, building a response model is not enough. You need to consider building a risk model. The most desperate applicants are often time the most risky.
- Firm Offer Of Credit, Always. Always use firm offer of credit and not ITA (invitation to apply). With a firm offer you can use languages that are direct and draw people’s attention.
- A/B Testing. For a variety of reasons, you might have opted out A/B testing. It’s not too late. Subtle differences in wording or color variances could lead to a significant increase in response rate.
- Landing Pages. If you haven’t build any landing pages, it could be hurting your pull through rate. To establish trust with your applicants, try aligning your direct mail creatives with the look and feel of your landing page. Also, have a look at your ads in google or bing, the language in your ad copy should also be as consistent as your CTA (Call to Action) on the mail piece.
- Re-mail Strategy. If you haven’t mailed the population that you’ve mailed before, you should. Try a different offer or a different creative. The amount of responders will be no less than the initial mailing. Remember, you never know when your customers will need your product or service.
- Know When to Pull Back. Your product offering may or may not work during a certain time of the year. Consider pulling back during tax return season and pump up the volume leading up to the holidays if you are offering a credit product.
- Use Multiple Mail Shops – Different shops have different core competencies. Some shops are great at fulfillment and shipping, some are great at printing. You might see a pricing difference, and a performance gain.
Bonus: Always test outside of your boundaries. Reserve 5% of your volume for these out of bound testing. Otherwise data you use to calibrate your program or subsequent models will become a self-fulfilling prophecy and die a slow death.
Check out my other articles: