01 · Unified customer behavior
Website, application, CRM, transaction, and campaign data contributed to a usable customer view.
Customer work · Retail and consumer
Ferns N Petals increased revenue by 3% with behavior-based personalization
Ferns N Petals had customer activity distributed across its website, app, CRM, transactions, and campaign systems. BluePi built an AWS customer data platform that unified those signals and supported timely behavior-based personalization.
Customer data platform for FNP personalization
FNP needed to reduce generic campaigns by using customer behavior already available across web, app, CRM, transaction, and campaign systems. BluePi built a customer data and personalization path with AWS Lambda, EMR, Glue, S3, RDS, Kinesis, and Redis to process events and support behavior-based campaign decisions.
02 · Timely campaign signals
Streaming and low-latency stores supported decisions while customer behavior still had commercial relevance.
03 · Measured personalization effect
The program delivered a 3 percent revenue increase, improved campaign efficiency, and made communication more relevant to each customer.
A customer profile path from events to campaigns
Customer activity entered through the data path, AWS services processed events and stored profile inputs, and the personalization layer used behavior to guide campaign decisions. Revenue, response, customer fatigue, and targeting efficiency connected the decisions to outcomes.
Unify the behavior record
Website, application, CRM, transaction, and campaign data contributed to a usable customer view.
Act on timely signals
Streaming and low-latency stores supported decisions while the behavior still had commercial relevance.
Measure campaign effect
Revenue, response, customer fatigue, and targeting efficiency connected interventions to outcomes.
Where this pattern fits
This case is useful when campaign volume is high but customer context is scattered. BluePi can start with one behavior, one segment, and one measurable intervention.
Working-session promptsQuestions that define the delivery boundary
Use these prompts to decide whether the case fits your operating problem and what a first deployment should prove.
- Which customer sources hold behavior that campaigns cannot see today?
- How quickly does a customer action need to reach a profile before it loses commercial value?
- Which campaign decision should a consolidated customer profile change first?
- How does the team measure response, revenue, and customer fatigue for each intervention?
- Who owns identity rules, consent boundaries, and profile updates after launch?
A practical first step
A useful first step is to pick one high-volume behavior, one segment, and one measurable campaign decision. BluePi maps the sources that describe that behavior, the freshness each decision needs, and the measure that proves value, then builds the smallest working path that connects them.
Case details
Open a section to review the customer problem, implementation, business change, and architecture.
01The starting point
FNP customer activity was spread across the website, mobile app, customer relationship management (CRM) system, transactions, and campaign records. No complete customer view existed, so campaigns stayed generic. The repetition caused customer fatigue and weakened campaign results.
- Siloed customer data: Customer data spread across the website, app, CRM, and other channels prevented a complete customer view.
- Limited personalization: Without a view of customer behavior across touchpoints, personalized marketing and engagement stayed out of reach.
- High noise to signal ratio: Generic campaigns caused customer fatigue and weakened campaign results.
02System delivered
BluePi built a customer data platform (CDP) on AWS that brings activity together as events arrive. Amazon Kinesis streams customer activity into the platform, AWS Lambda processes events in real time, and AWS Glue catalogs and prepares the data. Amazon S3 stores the consolidated record, Amazon EMR processes large datasets for analysis, Amazon RDS manages structured customer and transaction data, and Amazon ElastiCache for Redis caches profiles in memory so personalization lookups stay fast.
- AWS Lambda: Processes customer activity events in real time.
- Amazon EMR: Processes large customer datasets for analysis.
- AWS Glue: Catalogs and prepares data so teams have unified access.
- Amazon S3: Stores the consolidated customer data at scale.
- Amazon RDS: Manages structured customer and transaction data.
- Amazon Kinesis: Streams customer activity into the platform continuously.
- Amazon ElastiCache for Redis: Caches profiles in memory for fast personalization lookups.
03Outcomes
Personalized campaigns based on customer behavior lifted revenue by 3 percent. Targeting became more efficient, marketing costs came down, and customers received fewer irrelevant messages.
- 3% revenue increase: Personalized campaigns based on customer behavior lifted revenue by 3%.
- Improved customer experience: Relevant, timely communication reduced noise and improved customer satisfaction.
- Increased marketing efficiency: Better targeting improved campaign efficiency and reduced marketing costs.
- Improved decision-making: Teams make decisions from a complete picture of customer behavior.
04Constraints that shaped the design
A personalization decision has a short useful life. The platform had to act on behavior while it still carried commercial relevance, which pushed event processing and profile lookups toward streaming and in-memory serving. Every channel also had to work from the same customer record, or the fatigue problem would move between channels instead of ending.
05What changed for marketing
Marketing teams stopped assembling customer context channel by channel. Campaign decisions now start from a consolidated view of behavior, and teams measure revenue, response, and customer fatigue against each intervention rather than judging sends by volume.