ZELLEHow Zelle connected advertising to 115 million incremental transactions
Zelle is a digital payments network available through participating bank and credit union apps. Owned by seven banks, it allows people to send money directly between bank accounts.
Its growth depended on transaction volume. More Zelle transactions would demonstrate value to its banking partners, while giving customers another reason to use their banks’ apps.
The ChallengEZelle knew its advertising was building awareness. It could not determine whether that awareness led to more transactions.
The customer action happened inside its banking partners’ apps. Zelle could not place a pixel at the point of transaction or use standard attribution to connect media exposure with usage.
Prior modeling attempts had not solved the problem. The relationship between media and transactions could be delayed. Payment patterns around the first and fifteenth of the month affected activity. Zelle’s adoption was growing quickly, new banks were joining the network, and recurring transactions needed to be controlled for.
Zelle needed to know:
Whether advertising contributed to transactions
Which channels and campaigns drove consideration and transactions
How much to invest and how to allocate media dollars across channel
How long the effects lasted
THE INSIGHTThe right model depended on the right design
On the surface, the outcome appeared to be transactions. But the model also needed the context surrounding those transactions. That meant defining what Zelle wanted to learn before asking its data team for the inputs.
A Bayesian media mix model could use daily data and account for a business whose adoption was still changing. Connecting the model with Zelle’s brand consideration tracking would allow each channel to be evaluated against both consideration and transactions.
WHAT CHANGEDA model connecting media investment to consideration and transactions
Eden first defined the business problem and translated it into a specific data request. This created a bridge between the creative team and Zelle’s data team, giving the data team a clear view of what was needed and what the analysis needed to produce.
Zelle provided anonymized daily transaction data from its banking partners. Eden helped select the modeling partner, define the inputs, and bring critical business context into the model, including recurring payment patterns and new banks joining the network. The MMM showed what was happening, and Eden’s work helped explain the why.
The model connected media investment to:
Transaction contribution
Minimum spend thresholds
The time between exposure and transaction
The duration of media effects
Consideration and transactions by channel
This gave Zelle’s marketing and finance teams a shared view of what each part of the media mix was producing.
THE RESULTSThe model attributed 115 million incremental transactions to marketing
Between 2020 and 2022, the model attributed 115 million incremental transactions to marketing. Eden then used the model data to explain the conditions behind that contribution and turn the findings into media decisions.
The analysis showed that transaction contribution increased during major campaign periods, media needed to clear a minimum spend threshold, and campaign effects lasted for eight weeks. Video drove the majority of transactions, while social was the most efficient channel.
Zelle used these findings to set investment levels, balance channels, plan the time between campaigns, and evaluate video beyond its short-term impact. The company could now show its banking partners how advertising contributed to usage and make clearer decisions about what to do next.