When it comes to leveraging predictive analytics for your business, data security isn't just important - it's fundamental.
At Pecan, we understand that your data is one of your most valuable assets, which is why we've built our platform with security at its core. Your data remains exclusively yours and is never shared with any third parties. This comprehensive guide answers your most important questions about sharing data with Pecan safely and effectively. For detailed information about our comprehensive security measures, visit our information security page.
Is it safe to share my data with Pecan?
100%. Data sharing with Pecan is safe and secure. See our Privacy Policy.
How can I transfer my data to Pecan?
Upload a file (we support CSV and Excel formats).
Do I have enough data to work with Pecan?
The minimum data size is 1,000 entities. However, more data will usually produce more accurate models.
What historical depth should my data have?
Share historical data that extends back about twice as long as the future for which you’re trying to predict. Additionally, we typically recommend that you provide at least six months’ worth of data.
Which specific data tables do I need to import?
Provide a transactional table or raw activity table that represents the historical records of entities and the action you wish to predict, such as all the historical transactions of all customers, or all the purchases of all of your products.
In addition, you can add attribute tables, which are columns of additional data that characterize the activity and/or the object. These could include the number and types of products in a purchase, the cost, the type of payment, the customer profile (age, gender, etc.), and other kinds of information.
Pro Tip: We recommend you share transactional data in which each event includes a timestamp, rather than aggregated data.
Example use cases and required data
Demand Forecasting
Example predictive question:
Predict the number of units ordered per SKU by each customer in the next 2 months.
Table name and description | Importance |
SKU-level Transactional History – Shipments Table or Order Table | Must have to start |
Customer Master Table (typically demographic data) | Important |
Product / SKU Master Data (product attributes/category) | Important |
Daily SKU-Level Inventory Data | Important |
Raw Material Prices | Nice to have |
Historical SKU-Level Prices | Nice to have |
External Data Enrichment (e.g., macro trends, weather data, demographic data, etc.) | “Cool” to have |
Sales Forecasting
Example predictive question:
Predict the number of units shipped per SKU in the next 2 months.
Table name and description | Importance |
SKU-level Transactional History – Shipments Table or Order Table | Must have to start |
Customer Master Table (typically demographic data) | Important |
Product / SKU Master Data (product attributes/category) | Important |
Daily SKU-Level Inventory Data | Important |
Raw Material Prices | Nice to have |
Historical SKU-Level Prices | Nice to have |
External Data Enrichment (e.g., macro trends, weather data, demographic data, etc.) | Nice to have |
Churn - Reduction in Activity or Engagement Use Case
Example predictive question:
For every customer who made over $100 of purchases in the last year, predict the likelihood that they will spend less than $10 in the next month.
Table name and description | Importance |
Customer purchasing history(purchased product(s), quote, sales channel) | Must have to start |
Customer engagement data(website, app, call center, marketing campaigns) | Important |
Customer demographics | Important |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
Churn - Subscription Cancellation
Example predictive question:
For every customer with an active subscription for Product X in the last 2 months, predict the likelihood that they will cancel their subscription in the next 30 days.
Table name and description | Importance |
Subscription Table(all active and inactive subscriptions including historical) | Must have to start |
Customer purchasing history(purchased products, quote, sales channel) | Important |
Customer engagement data(app, website, call-center, marketing campaigns) | Important |
Customer demographics | Important |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
Lifetime Value
Example predictive question:
Predict, 7 days from registration, the user’s lifetime value on day 90.
Table name and description | Importance |
Customer purchasing history(purchased product, quote, sales channel) | Must have to start |
Customer engagement data | Must have to start |
Customer demographics | Nice to have |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
Competitor new product data | Nice to have |
Competitive/modular/pricing strategy inputs | Nice to have |
Conversion
Example predictive question:
Predict, 2 days from registration, the likelihood of a user to convert by day 30.
Table name and description | Importance |
Customer purchasing history(purchased products, quote, sales channel) | Must have to start |
Customer engagement data | Must have to start |
Customer demographics | Nice to have |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
Upsell
Example predictive question:
Predict, for every customer who made a purchase in the last 6 months, the likelihood that they will make another purchase in the next 2 months if called or offered a discount.
Table name and description | Importance |
Customer purchasing history(purchased products, quote, sales channel) | Must have to start |
Customer engagement data | Must have to start |
Customer demographics | Nice to have |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
High Value
Example predictive question:
Predict, 2 days from registration, the likelihood of a user becoming high-value (paying more than $500).
Table name and description | Importance |
Customer purchasing history (purchased products, quote, sales channel) | Must have to start |
Customer engagement data | Must have to start |
Customer demographics | Nice to have |
Product/SKU master data (product attributes/category) | Nice to have |
Pricing/promotion strategy | Nice to have |
Lead Scoring
Example predictive question:
Predict the likelihood of an inbound lead to convert to “Closed Won” if approached by an AE or BDR.
Table name and description | Importance |
Historical Pipeline Data (leads information) | Must have to start |
Customer Purchasing History / Closed Deal Table(Joinable to Leads table) | Must have to start |
Any Available Lead Information(demographic, third-party company information, marketing engagement, forms filled) | Nice to have |
Sales Rep Data | Nice to have |