Deep personalization means offering the right products in the right lifecycle context on any marketing channel. The right offers are appreciated and motivate purchases. The wrong offers are irrelevant, annoying, and increasingly ignored. Segmentation is a blunt instrument for the job. The way to get it right is to analyze and score at the individual customer level.
Automated, Self-Learning Analytics Service
Purchases(Transaction Data)
Predictive Scores by Customer, Campaign Lists, Product Recommendations, Reports
We know from experience that the best predictor of future purchasing behavior is past purchasing behavior, so we use the always available, always complete, never ambiguous, transaction data you already have on every customer. We take a long range view for deep personalization, usually transactions over 3-5 years, and learn continuously as new data is updated.
We provide metrics and scores for each customer in any format you need, whether to drive marketing automation systems, inform recommendation engines and CRM systems, enrich customer databases, or add new dimensions to reports. In addition to your data the way you want it, we provide a standard set of informative reports on the state of your customer base with each analysis.
In addition to customer loyalty predictions, we score every customer on likelihood to buy every product in one pass. A few thousand customers or many millions, it does not matter. A few dozen products or hundreds of thousands, no problem.
Originating from research at the University of New Hampshire, and refined by the experience of many hundreds of marketing campaigns, our PSL Machine is a unique, proprietary blend of many techniques, including survival models, modified Markov chains (with memory), machine learning, and dual lattice arrays (for every customer and every product) where probabilistic inter-dependencies between derived variables generate state change predictions.
See a list of predictive metrics generated for each customerBy mining the purchasing patterns found across the entire customer base, we build a Probabilistic State Model for every individual customer automatically. The models are automatically updated when new transaction data becomes available. The approach is specifically designed not to over-complicate models or introduce weak, stale, or negative predictor variables. The result is lifecycle stage, value, risk, and product purchasing predictions you can rely on to target audiences and personalize messages and offers that engage and motivate.
| Analytic Approach for Deep Personalization | Analytics Platform | Some Marketing Platforms | Loyalty Builders Service |
|---|---|---|---|
| No platform to setup, maintain | |||
| Used by analytics experts for a wide range of problems | |||
| Shows offers based on customer segment or product browsing | |||
| Provides best 1:1 offers any time, regardless of recent behavior | |||
| Deep analysis of all customers and all products at once | |||
| Uses always available, reliable data on every customer | |||
| No customer profiling or databases to maintain | |||
| No third party data or sensitive personal data | |||
| No integration of data across different sources | |||
| No care or maintenance of models, rules, etc. |
While marketing platforms have so far automated how and when to send messages to customers, Loyalty Builders solves the "last mile" by automating what to send inside those messages — ready at any time with the best offers for each or every customer to get them buying.
