Great customer data platform with even better UI and big data capabilities...
July 27, 2020
Great customer data platform with even better UI and big data capabilities...
Score 9 out of 10
Vetted Review
Verified User
Overall Satisfaction with Treasure Data
Treasure Data is used as the CDP for one of the regions (APEC) primarily for creating master customer profiles. Currently, we are using it for various use cases like Abandoned Cart/Abandoned Browse, Social Media campaigns, RFM segments, Cross-sell, etc. Additionally, the master segment and reporting/dashboarding functionalities are being used to track the growth of the number of customer profiles, attribution by various email sign-up sources, other reporting uses cases.
I have found the step-by-step workflow in building multiple segments very useful, which can be either run on-demand or scheduled for automated execution and campaign activations. The next steps are to leverage machine learning models for additional use cases like predictive scoring, etc.
I have found the step-by-step workflow in building multiple segments very useful, which can be either run on-demand or scheduled for automated execution and campaign activations. The next steps are to leverage machine learning models for additional use cases like predictive scoring, etc.
- Providing omni-channel view of customer behavior
- Data ingestion including batch and near real-time data
- Campaign activation including email and social media
- Great UI, Flexible Data Model to work with given enterprise data
- Documentation could be better for the workflows & any custom logic implemented. Seems to be improving with the new approach using Markdown.
- API key management can be improved further
- Better customer engagement levels
- Reducing churn and increasing customer LTV
- Campaigns are now more effective resulting in higher RPU (revenue per user), better CTO rates, etc.
- Calculating the inferred product gender
- Leveraging the PySpark API to get the curated/profiled data back from Treasure Data for the purpose of integrating with the data lake
- Agilone and Acquia Platform
Based on my experience, the most striking difference between the two platforms are the way their data models are organized.
Agilone (now part of Acquia) has a very hard/strict requirement for integration with the source systems as we need to conform/adhere to their canonical/existing data model while Treasure data is quite flexible in being able to work with the respective source data as long as the relevant data was present. Overall, this has translated into a faster time to market in the case of Treasure Data and flexibility in incorporating changes/enhancements.
Agilone (now part of Acquia) has a very hard/strict requirement for integration with the source systems as we need to conform/adhere to their canonical/existing data model while Treasure data is quite flexible in being able to work with the respective source data as long as the relevant data was present. Overall, this has translated into a faster time to market in the case of Treasure Data and flexibility in incorporating changes/enhancements.
Using Treasure Data
12 - Treasure Data is mostly used within our eCommerce and Digital Marketing, Data Science and Data Analytics, and CRM (Customer Relationship Management) functions.
6 - To support Treasure Data integrations, the team has skills related to Data Engineering (Azure Data Factory, Databricks, Apache Spark, ADLS, etc.) and Data Science (Python, ML, etc.)
- Customer Data Platform for maintaining the 360-degree view of the customer
- Activating email and social media campaigns
- Tracking RFM scores & increasing Customer LTVs
- Enhanced reporting and dashboarding
- Building propensity models
- Better retention of lapsing guests and win back of churned guests
Treasure Data Support
Pros | Cons |
---|---|
Quick Resolution Good followup Knowledgeable team Problems get solved Kept well informed No escalation required Immediate help available Support understands my problem Support cares about my success Quick Initial Response | None |
Yes
We require the support as we are still in the process of enhancing and building new uses for the business functions like building propensity models based on various criteria like propensity to buy on markdowns/discounts, propensity to buy certain product categories, etc. There is a backlog of features like the above as well as integration of new geographies that are planned in the near future that requires further support.
Yes - The Pyspark API was not working with datasets that have array columns/datatypes. This was promptly resolved upon reporting to their technical team who in turn raised it with their product team.
When we were faced with a couple of issues in being able to integrate the curated data from Arm Treasure Data platform back to our Azure DataLake, we reached out to Arm Treasure Data's support team. They provided the much-needed support on the changes to be made in the scripts and the right way to apply the API keys to successfully establish the connection and resolve the issue.