Prediction and purchase dashboard

Reduce the risk of losing a client by predicting his actions based on collected transactional and behavioural data. By using Big Data, you can quickly and effectively respond to a decrease in customer interest in your product by adjusting your marketing activities accordingly. Predictive analysis will allow you to estimate the risk of customers churn and the likelihood of making another purchase.


CONTENTS

  1. Customer Lifetime Value Analysis
  2. Prediction value of purchase occurrence
  3. Prediction value of the transaction from source occurrence
  4. Churn prediction value

To go to the dashboard of purchase prediction and churn analysis, go to Artificial Intelligence → Predictive Analysis 


1. CUSTOMER LIFETIME VALUE ANALYSIS

The data presented are based on the whole database of contacts, which are contained in the client’s account. This data are based on the historical (until now) behaviour of contacts in the database.

[1] Average customer lifetime value the average amount for which a contact makes a purchase while “being an active customer / active contact in the database”. 

[2] Average time between customer purchases – the average time between customer purchases 

[3] Average number of customer purchases – i.e. how much one customer on average make a purchase before he or she completely gives up our services

[4] Average time since last customer purchase how much time has passed on average since last purchase for one contact 

[5] Average client’s lifetime – how much time passes between the first and last client purchase 

[6] The average age of a customer – how much time passes between the first purchase and now.


2. PREDICTION VALUE OF PURCHASE OCCURENCE

Presents the information resulting from the analysis of all customer data since 2018. Therefore, the size of the sample may be greater than the current number of contacts in the database. On the basis of all historical data about contacts (activity on the website, making purchases), the contact database is analyzed and the average value of purchase prediction for the examined sample is determined. 

[1] Sample size – the number of clients from the contact database ( since 2018) on which the research has been conducted

[2] Average value of purchase prediction for the sample – average value for a given group of contacts, determining the probability of making a purchase again 

[3] The legend of values

[4] Diagram showing the estimated value of purchase prediction for a given sample


3. PREDICTION VALUE OF PURCHASE FROM-SOURCE OCCURRENCE

The prediction of purchase occurrence module, presents estimated data, determining the probability, with an indication of the source that brings us the greatest probability of completing the purchase by the contact. The data is calculated on the basis of transactional data of a given contact and its visits to the website. The module has been divided into 4 ranges, with an indication of the probability value.

[1] Recommendation frame [clicked] –  making a purchase by clicking on a product in the recommendation frame 

[2] Referral – making a purchase after entering the website from another website; 

[3] Advert – making a purchase after entering the website by clicking on Google Ads;

[4] Search – making a purchase after entering the website by clicking on the link resulting from the search (in the search engine);

[5] Web Push from rule [clicked] – making a purchase after entering the website by clicking on the Web Push notification sent using automation rules;

[6] Mass SMS [sent] – making a purchase by entering the website after receiving a mass SMS;

[7] Mass Web Push [clicked] – making a purchase after entering the website by  clicking on the mass Web Push notification;

[8] Email, as a rule [clicked] – making a purchase after entering the website by clicking on the email sent from the automation rules;

[9] UTM – making a purchase by entering the website from the link containing the UTM parameter;

[10] Mass email [clicked] – making a purchase after entering the website by clicking on an email sent in bulk;

[11] Web Push from rule [opened] – making a purchase by entering the website after opening the Web Push notification sent using automation rules;

[12] SMS usually [sent] – making a purchase by entering the website after receiving an SMS sent by means of automation rules;

[13] SMS from Workflow [sent] – making a purchase after receiving an SMS sent from Workflow;

[14] Workflow email [click] – making a purchase after clicking on the email sent from Workflow;

[15] Email – making a purchase after entering an email;

[16] Legend – the percentage value’s meaning in the context of the probability of a transaction occurring


4. CHURN PREDICTION VALUE

It presents an analysis of all client data since 2018. Therefore, the size of the sample may be greater than the current number of contacts in the database. The value of the contact churn prediction is determined on the basis of historical data, i.e. it shows what part of the tested sample will not make a transaction again. The data is calculated based on historical purchases made by contact. 

[1] Sample size 

[2] Average churn prediction value for the sample

[3] Legend

[4] Diagram showing the estimated value of customer churn prediction for a given sample

If you need more information about the topic mentioned above, please contact us: support@salesmanago.com +1 800 960 0640