While most brands still make online ad buys based on cookies and bidstream data evaluated in milliseconds, publishers are starting to accommodate forward-thinking advertisers.
For instance, Angie’s List senior digital marketing director Keary Phillips is helping to bring in offline data to inform real-time bidding.
And Google and Facebook allow brands to attach a cost per acquisition (CPA) to customer IDs, which the platforms use to optimize toward high-LTV potential customers.
“Facebook and Google have conversion optimizer technology where you can give a CPA target and your data and let them go do the hunting for you,” said Jess Jacobs, marketing director at the online retailer Wayfair.
But applying the same tactics on open exchanges can be a challenge.
Instead of relying on the walled gardens, Wayfair opted to build its own optimization and predictive LTV products on top of AppNexus, Jacobs said.
“In the past couple months,” she said, “we’ve been able to directly plug our model that predicts incremental value from an ad directly to AppNexus so we don’t have to simplify it in order to buy on the exchange.”
So why is it so hard to apply LTV metrics to the open exchange?
According to Artem Mariychin, co-founder and CEO of the LTV analytics firm Zodiac, RTB buyers tend to value inventory based not on offline data, but on engagement or impressions.
Applying LTV metrics could make an otherwise successful campaign look bad. Consider a company like Blue Apron: the meal-delivery service loses money on about 70% of the customers it acquires due to churn before they return profit, according to a model from Zodiac based on Blue Apron’s IPO disclosure data.
“For a brand like Blue Apron that’s paying $90 or more for a new user,” Mariychin said, “an ostensibly successful campaign that drove lots of traffic and sign-ups but ended up retaining few loyal customers can be punishing.”
Blue Apron declined a request for comment.
Also, agencies and vendors don’t want to wait for months to confirm retention before getting paid for their labor. Some demand-side players are trying to build products around LTV estimates, but few have the deep working relationships necessary to incorporate a brand’s data modeling.
As a compromise, agencies and DSPs flatten out customer data into lists that reflect average value, or split out campaigns for high-value or low-value targets, said Ari Paparo, founder and CEO of the bidding tech provider Beeswax.
“Marketers want to bring real-time scoring to each ad impression instead of average scoring,” Paparo said. “But this is technically complex since the data needs to be closely integrated into the DSP’s bidding system.”
DSPs are beginning to build tools that deliver returns within a brand’s CRM, not just on the campaign media report, Phillips said. “We’re asking for that capability from our partners and vendors, and I think we’re just at the cusp of that from a marketer standpoint.”
For Wayfair, it made more sense to begin the journey on its own.
Wayfair’s first incremental return after adopting its dynamic CPM strategy was “being able to more finely tune our bidding down on user level, taking into account lifetime value,” Jacobs said. “Facebook and Google have their conversion optimizer technology, but we found we can get a lot more juice from the squeeze by taking control of that ourselves.”
This post was syndicated from Ad Exchanger.