Introduction
Enterprises that have undergone a digital transformation, or who are planning one, understand that their data is their most valuable asset. Most enterprises list bad customer experience as their reason for leaving a service behind. It has become critical for brands to harness, harvest, and analyze their customer experience data.
For years, successful brands have been leveraging their data from their CRMs, websites, POS systems, surveys, and social media to better understand their customers. Unfortunately, many companies have overlooked the value of the authentic unified multichannel conversations between their customers, agents, and bots; the source of which could be various speech and text conversations like a phone (VoIP telephony), SMS texting, email, web chat, social media, video conferencing, etc. Due to the size and complexity of working with unstructured voice and text data, contact center interactions have proven difficult for brands to deeply analyze.
Enter Analytics Studio, by LivePerson, a leader in delivering intent-driven conversational experiences at scale, recently announced a groundbreaking new solution that enables data-driven enterprises to harness the power in their omnichannel call recordings and text conversations that until now have been hiding in plain sight. Analytics Studio is a tightly integrated, turn-key analytics platform that converts conversations into rich actionable data within minutes.
The product captures conversations through LivePerson’s Conversational Cloud where it can collate asynchronous messaging and synchronous chats, and also speech through LivePerson’s VoiceBase platform which has a powerful speech-to-text and speech analytics API where calls are transcribed.
The conversations are measured with metrics like sentiment, silence, survey results, MCS, and NPS and tagged with customizable key events like conversation drivers, and competitor mentions. The data from each conversation is combined with the metadata of their customers so that each conversation can be tagged with meaningful information about the customer’s, agent’s, and bot’s behavior impacting an interaction’s outcome. Next, through a simple online configuration, this rich data is transformed and offered back to the customer in a relational data schema that customers can host in their own Snowflake, RedShift, Exasol, or other data management solutions. This approach provides clients with complete control and ownership of their data. Analytics Studio pairs all of this data with a complete suite of detailed Tableau dashboards which can be customized to fit the individual needs of each customer.
To increase speed to insight, Analytics Studio also offers a hosted version; because we understand that constraints on resources and priorities can often become bottlenecks to even the most promising projects. With this option, LivePerson does the hard work for you by managing the data warehouse, Tableau Server, and dashboard maintenance. This allows business users to get immediate answers from their data and begin reaping the ROI rewards.
Analytics process
Either voice or textual conversations are analyzed to match the different intents of a conversation through categorization (see VoiceBase). These intents are then used to determine the effectiveness of agents to handle consumers, the needs of the consumer, necessary product enhancements, etc.
Analytics Studio is part of the LivePerson Analytics Suite of applications and has built in integration with the LivePerson Messaging Interactions API (MIA) to analyze and report on Conversational Cloud Messaging.
Basic architecture
Architectural components
Conversation Integration
- Call recordings: Audio recordings in most media types are posted to the audio endpoint for transcription, analysis, and reporting.
- Metadata: Additional information outside of the conversations, such as consumer/visitor ID, phone number, agent ID/name, etc. should be attached to the conversation. LivePerson MIA metadata is automatically attached.
- LivePerson Messaging: The LivePerson MIA API (liveperson.net) is fully integrated with Analytics Studio, to seamlessly acquire Conversation Cloud conversations. No need to map conversational metadata using this API, since LivePerson MIA metadata is automatically attached. The most recently completed conversations are read every 10 minutes to keep analytics report data fresh.
- Rest API text endpoint: Other third-party messaging conversations are integrated with Analytics Studio.
Transcript and analytics engine
App and API platform: Each conversation is processed individually using VBQL Categorization, LLM, Machine Learning, or simple keyword searches to identify and categorize what is spoken or typed between agents and consumers. Audio media is transcribed into text before processing. The results for each conversation are returned either through efficient callbacks, or API calls, or batched together using the ETL process. This process is run about every 30+ minutes to balance combining large batches of conversations with keeping analytics results fresh. Larger batches ensure high data load efficiencies. Under higher loads where higher batch sizes can be maintained, loads will occur more frequently.
ETL process
Conversation results are analyzed and combined into tabular, aggregated results ready for the Data Lake. This process is responsible for batching thousands of conversations together for a maximum period of time, to make downstream processing more straightforward. This includes creating a SQL-style, normalized, tabular format ready for importing. Typically these results are hosted in LivePerson’s own Exasol in-memory Data Warehouse, and queried using Tableau. However, this data is also available to be exported into external Data Warehouses in a parquet format as well.
Data warehouse
Exasol: LivePerson uses Exasol, an in-memory SQL database to store our brand’s hosted conversation data which is ready for generating analytics dashboards and reports. We have reserved more than 2TB or memory across a cluster of hardware to ensure solid performance.
Dashboards
Tableau: Out of the box shared dashboards, as well as brand specific custom reports are developed and viewed from a Tableau Web Server framework. Each brand’s data is kept in a schema with separate credentials (schema-based access control) to enhance security.
LivePerson integration diagram
FAQs
- Does LivePerson resell Tableau?
No, LivePerson is not a Tableau reseller. The license used by customers is not available to be used on other tableau servers or for other uses. - Does Analytics Studio work with other BI tools?
Yes, we have focused on Tableau and use that internally for Hosted Analytics Studio but the same data can be visualized in ThoughtSpot, PowerBI, and Qlik. We plan to create ready-made templates in these tools soon. - Does Analytics Studio work with any data warehouse/database?
Analytics Studio works well with most data warehouses/databases. However, like any other analytical system, it works well with columnar storage systems like Exasol, Redshift, etc. As a thumb-rule, we recommend using any MPP, columnar system for optimal performance. - What integrations does Analytics Studio have out of the box?
- LP Conversation Cloud Messaging History API
- LP Conversation Cloud Engagement History API
- LP’s VoiceBase Voice Media API (REST API)
- LP’s VoiceBase Text API (REST API)
- Amazon S3
- Azure Blob (Q3 2023)
- How do customers connect conversations if they don’t have one of those integrations?
Most of our customers today send text their conversations through Conversational Cloud or have built an API integration to send conversations directly to our REST API. LP Customer Success team works with customers to support their integration needs for this. - What metrics would we get access to through Analytics Studio?
We offer>50 conversation metrics on text conversations and on stereo call recordings including sentiment, response time, survey analytics, repeat contacts, CSAT, NPS, MCS, etc. Some of the metrics are specific to conversation types (text vs. speech). Please schedule a demo with us to know more. - How fast is the turnaround time to get the conversation analytics results?
Typically within 20 minutes from the time a conversation is hooked up to Analytics Studio, the related analytic data is published to the Tableau dashboard. - Does Analytics Studio support SSO
Yes, although Analytics Studio is based on LivePerson’s VoiceBase platform and cannot yet be accessed through Conversational Cloud, we do support SSO integration and can work with customers to enable the same. In the near future, we plan to have a unified interface for Analytics Studio through Conversational Cloud. - What do current LP Insights customers lose?
The current Insights customers lose none of the features. On the contrary, customers will get access to better interfaces, and a plethora of features that have been listed above. - Is there any action to be taken for the existing LP Insights customers?
LivePerson will work with the account managers and migrate your account to the new product - all accounts will be migrated by the end of Q3 2023. Account managers will reach out to customers with the amendment to the LivePerson order form that needs to be signed. - What effect does this have on the existing VoiceBase customers?
VoiceBase is now being integrated as a part of the LivePerson family after the acquisition. Nothing changes for VB customers using Enterprise Accelerator (EA), the same product with new LP branding and a new name, plus an ability to analyze messages along with voice. A simplified, more intuitive UI that offers easier navigation while elevating your data capture to span both voice and messaging analytics. The new UI with LP branding would be visible from August 1st, 2023. - What is the pricing for Analytics Studio?
Please reach out to your LivePerson account representative for the details.