What if I Could Make Your Remarketing Advertising Smarter?
If you live and die by the website, you know getting a second bite of the apple is critical. Remarketing works assuming that you have good quality ads, you have the right network, etc.; but how do you get it to work better? What if I told you I could make your remarketing ad campaigns smarter?
At the end of the day, even a 1% increase over your conversion rates has a big impact on the business. Even though you are optimized, it may be time to look at some sophisticated machine learning to boost your conversion rates.
There are two main areas that we think could have major impact on driving higher remarketing conversion rates:
- Identify shoppers from browsers – some of that you can tell by what page they left your site, but some of it is opaque. We don’t have enough detail. But, what if you could analyze how they engaged and make some determination as to their likelihood to buy?
- Where and how we target ad impressions – we serve ads to everyone we can identify to see what sticks – What if you could target based upon profiles of known buyers, different ads to different people at different points based upon their identified behavior patterns?
What would you get to improve your remarketing?
- Likely Buyer Score – High score gets more follow up as they look like real shoppers, low scores get light footprint as they look like simple browsers.
- Dynamic Clustering – target profiles with ads that you know will work based upon recognition of their behavior.
How much work does it take? A lot of work. No, really we are just kidding. It is very simple.
- Simple code on your website
- Export of the score and cluster information to your systems
- Plugin to your remarketing efforts
- Test efficacy, repeat and rinse to optimize
The science behind the scoring is pretty sophisticated machine learning, but the application for juicing your remarketing efforts is really simple. Plugin a score, refine your ad serving, and reap the rewards. The beauty of this model is that it is easy to A/B test the results. Numbers speak for themselves.
You can read the full article about machine automation here.