JD Cloud Zhike Marketing Solutions

Fully explore users’ offline and online behaviors, precisely position target customer groups, provide merchants with services such as site selection and investment, marketing and promotion/operation, customer group insight

Product Architecture

Zhike Product Architecture

In terms of Data Source, support for offline structured data and real-time data. In terms of Data Integration, support for data gateway, use Flume collection log, Kafka collection log, Sqoop offline access, Gobblin as ETL tools. In terms of Data Storage, use HDFS as distributed file system, use HBase as distributed database. In terms of Resource Management, use Yarn for unified resource management and scheduling. As for data computation, off-line computation is realized with Hive, MapReduce and Spark and real-time computing is realized through Spark Streaming. As for job scheduling, Oozie serves as the job stream scheduling system.

Business Scenarios

Zhike Business Scenarios

User insight: integrate customer's own data assets and JD data, analyze customer’s behavior, preferences, and basic characteristics of the whole network. Digital marketing: Extract population precisely according to behavioral and attribute characteristics; perform precise marketing by using dsp tools and SMS channels owned by JD, in combination with a variety of cooperation channels. Marketing effect evaluation: Integrate the monitoring data provided by brand customers for the whole network delivery with the data of users' multiple behavioral dimensions of JD Mall in reliance with big data calculation capability to evaluate advertisement effects.

Main Services

Crowd Profiling

Typical Scenario: 1. Member profiling: Integrate the member information provided by the enterprise with JD profiling through being transmitted through the API, uploaded locally to output the profiling 2. Customized user profiling: Provide crowd information collected by crowd activities (support local upload), integrate with JD profiling to output the profiling; generate customized profiling for the crowd extracted through a variety of online behavior combinations. 3. Crowd screening: Use cross combinations of existing tag dimensions (up to three tags) to screen the crowd

Solution: Integrate JD data with full data from store collection, customer upload, analyze the multi-dimensional data of customer’s whole-network channel behavior and preference, group basic characteristics and interest preference to analyze customer characteristics, membership stratification and customer loyalty for merchants to help merchants fully improve the marketing capability.

Use Now

Digital marketing

Typical Scenario: 1. Precise establishment of crowd package: Establish the crowd package, local upload crowd package through profiling, extract crowd package through customized combination of online shopping mall behaviors. Provide BlackDragon and JZT with JD’s own advertisement delivery system identification, in which the advertisement can be delivered in its own online store APP, etc. 2. DSP delivery: Select precise crowd to deliver by JD internal promotion tools, in which the advertising space can be selected, the advertising space blacklist and whitelist can be established for managing advertising space. 3. SMS delivery: Use customized groups for SMS delivery, in which the delivery channels can be selected on its own, SMS templates can be created, and the SMS delivery time can be designed.

Solution: Improve the user identification policy to help fine management of target groups by brand customers. Through the channel advantage of JD promotion, expand the coverage of media delivery and help merchants to conduct precise marketing.

Use Now

Marketing Effect Evaluation

Typical Scenario: Marketing effect evaluation: Integrate the monitoring data for the whole network delivery with the JD data to provide advertising brand influence and reached data

Solution: The marketing environments and media touchpoints are increasingly complex so that the traditional marketing effect evaluation method based on consumer recall has failed. Brand advertising highly focuses on the promotion of brand sales. The computing capacity of JD Cloud Big Data provides brand customers with the monitoring data for the whole network delivery with the multidimensional data of user browse, search and purchase of JD Mall by the user to supplement the real user brand influence and touched data of advertising delivery to evaluate the advertising effects.

Use Now

Benefits

  • All-dimensional Data

    Multi-dimensional data comprehensively reflect the operation of shopping malls, business districts, shops, customers and activity marketing

  • Leading Data Algorithm

    United with USC Data Science Lab to have the leading data algorithm in the world

  • Intelligent Analysis and Decision

    Learn based on machine, analyze the site health status analysis for reminder, and guide the activity marketing

  • Light and Convenient Use

    The product is compatible with PC client and mobile phone client to view data with multi-level and multi-permissions

Case Studies

  • CloudXink

    Marketing Effect Evaluation

    CloudXink

    Provide the touched data of dimensions that the whole network advertising is transformed into browse, search and purchase by the user of JD Mall to evaluate the advertising effects

  • Second Hand

    Profiling, Marketing Delivery and Marketing Effect Evaluation

    Second Hand

    Provide multiple dimensions such as society, region, time, consumption shopping preferences to have a deep insight into the crowd; screen the crowd by behavior and static profiling for precise delivery

  • Jingpaidang

    Profiling, Marketing Delivery

    Jingpaidang

    Provide multiple dimensions such as society, region, time, consumption shopping preferences to have a deep insight into the crowd; screen the crowd by behavior and static profiling for precise delivery

  • Ipsos

    Marketing Effect Evaluation

    Ipsos

    Provide the touched data of dimensions that the whole network advertising is transformed into browse, search and purchase by the user of JD Mall to evaluate the advertising effects

  • iPinYou

    Marketing Effect Evaluation

    iPinYou

    Provide the touched data of dimensions that the whole network advertising is transformed into browse, search and purchase by the user of JD Mall to evaluate the advertising effects

  • Domob

    Marketing Effect Evaluation

    Domob

    Provide the touched data of dimensions that the whole network advertising is transformed into browse, search and purchase by the user of JD Mall to evaluate the advertising effects

  • 180.ai

    Marketing Effect Evaluation

    180.ai

    Provide the touched data of dimensions that the whole network advertising is transformed into browse, search and purchase by the user of JD Mall to evaluate the advertising effects