Interview with Jimmy Hu on CDP Technology — by AI Front
Big data has dramatically changed the traditional marketing approach that relies on brand communication. In 2018, Coca-Cola, with a 132-year history of strong marketing and promotional capabilities, announced the recision of the position of “Chief Marketing Officer” and introduced the new role of “Chief Growth Officer”. Traditional brands spend up to hundreds of millions of dollars each year on marketing alone. Marketing needs to be driven by data and technology to make it tailored towards the end consumer and achieve more quantifiable results.
“Internet Marketing” has dominated the past decade, and e-commerce together with mobile payments have changed consumer behavior and the way people live. In the next decade, online shops and offline entities will be integrated into the same operation, and the effective combined use of online and offline traffic can only be achieved through more refined and customer oriented operations. The data-driven and closed-loop marketing, the transformation of the traditional enterprise as represented by Coca-Cola, is spreading across industries.
Today we have a conversation with Jimmy Hu, CEO of APEX Technologies, a customer data technology and AI company headquartered in Shanghai with offices in Beijing, Hong Kong and Silicon Valley. Its flagship product is Nexus, an Enterprise Customer Data Platform. The team has grown from less than 10 people to nearly 200 people recently. Furthermore, APEX Technologies has also managed to rapidly expand their customer base in the past two years to 300+ large and medium-sized enterprises. In the interview, Jimmy Hu, founder & CEO of APEX Technologies, spoke of one of the reasons for the rapid growth: “The rise of big data marketing technology — There are some trends now, some of which have previously saved a lot of money in advertising or traditional marketing activities. Big brands, whether in FMCG or the automotive industry, are tightening budgets and reducing waste, using data to drive refined and more effective marketing operations.“
Seize the moment and move at the right time
The Customer Data Platform (CDP) is a cutting-edge data management system that brings together internal data and then integrates that into customer life cycle related scenarios (marketing, operations, customer service, etc.). Being in the limelight in North America, this marketing technology is well employed by companies with a market valuation of more than $1 billion. In China, on the contrary, businesses just started to adopt CDP technology and APEX Technologies is the leading player. A CDP will collect from each of the fragmented online and offline data sources, including retail outlets, APP, CRM, databases and other business related data sources and ultimately form a 360-degree customer portrait. This will facilitate tailored marketing based on customer life cycle insights and will enhance the customer experience.
In order to position ourselves strategically as the market adopts CDP technology we consider three important developments. The first is to complete the iterative development of the product, and launch a mature solution at the right time.
The second is that in the era of big data, the customer data platform also needs to take into account the amount of data accumulated in the past. Some leading enterprises already have some off-the-shelf, internal data collection methods. Starbucks is one of APEX’s customers, and it has been setting up an enterprise-class data platform long before, integrating all offline stores and online data as part of their digital transformation. With such big data foundation, the CDP platform was subsequently introduced. Based on AI product scenarios such as product recommendation in the APP, customer churn rate and customer value forecasting, etc., the company is able to take appropriate measures and adapt their business strategy. Because of the enormous disruptive impact of the Internet, the coffee industry leader who sits firmly in the top has to increase their technical marketing component to enhance their core competitiveness.
Lastly, the effects of using data transformation are measurable. Jimmy Hu mentioned another customer, Spring Airlines. With the initial route recommendation, the conversion rate of the push notification is only 1%. After several months of using CDP and additional AI modules, the direct conversion rate increased to 10% and later to 14%. The conversion rate has increased by more than 10%, and the economic benefits are tens of millions of RMB. This visible and tangible result is very convincing for industry leaders to think about enterprise transformation.
Just the right turn
When companies and their operations shift from traditional operations and management models to modern, technology-oriented operations, this transformation is known as digital transformation, which has become a necessity for all companies as it broadens their market. However, when the layout of the company started in recent years, the main focus is on five to six fields. Jimmy Hu said: “These vertical areas have their own characteristics. We are not doing the transformation for all B2C companies. The first batch of industries are financial services, automobile, tourism, new retail and education.”
The United States based company, General Electric (GE) has always been known as a role model for digital transformation of an enterprise. When the development was vigorously carried out until 2018, the media communicated the message of “transformation failure”. One of their most important products, “Predix”, has abandoned the ambitious plan to become the Industrial Internet of Things Android-like system and finally re-adjusted to only three areas: aviation, medical and energy systems.
Jimmy Hu added: “Our customer’s business environment is screened thoroughly before technology does its thing. For example, the customer’s data volume reaches a certain level, which requires refined operation and data-driven management to improve the customer repurchase rate and maintain the customer life cycle. There will also be different solutions according to these five or six segments. For example, the AI model is tuned according to the characteristics of different industries and different companies.”
How to make data more valuable
There are two core functions of CDP Technology. One is data acquisition, processing, integration, and analysis — basically making the data “applicable”, and the second is to extract the data value with AI. The products correspond to NEXUS and IQ. NEXUS provides a range of services from data collection, data cleaning, data correlation, labeling systems to data analysis. It is the current core platform product. IQ is an AI tool built on NEXUS. The data processed by NEXUS is centralized on IQ. The data source of IQ can be real-time, non-real-time, structured, unstructured data, or even textual data. IQ puts this data into a model and can implement purchase forecasting, product content recommendation, intelligent early consumer warning mechanisms and user life cycle prediction scenarios. There are currently more than 10 commonly used AI models implemented with our enterprise clients.
Among them is data processing, a very critical link. Data processing entails how to quickly import data from various data sources into CDP and get to a unified and usable format. This is where many competing products perform very weakly. NEXUS is very flexible in this regard as it can be directly imported into various data sources and converted into the required format. It is also possible to open and converge data from different systems. This convergence of two systems needs a team of dedicated people, or can be outsourced to a third-party. It may take 2–3 months to develop through API, and the data middleware platform may only take a week or less. Data processing can be done completely through the product interface, and can be deployed without the assistance of IT staff. This also avoids the problem that every project in the enterprise has to set up a corresponding team to operate and maintain the system.
The second step is the link to extract the value of the data after the data is processed. Using the AI module “IQ”, customers can directly call the pre-made basic algorithms and models in the module, or customize the relevant models, including customer grouping models, personalized recommendations, sentiment analysis, loss warning analysis models, customer value models, marketing models, scoring models, etc.
Data can flow in and out, and the incoming data always belongs to the enterprise itself, a feature of the CDP platform. On the contrary, some platforms of companies such as Ali’s, have become an internal data bank because these companies’ policies, traffic and data are the biggest realizable assets for them, so they can’t use an open data platform to work with external corporate customers. In other words, the data of the enterprise customer can only enter this system, but it cannot be released. This kind of product is not a platform that enterprise customers can control themselves, and enterprises do not want to continue to give their data assets to such data banks. APEX Technologies introduced CDP technology to integrate first-party customer data into a unified interface, and then, combined with such real-time behavior data as CRM data and third-party data, facilitates to create a complete user portrait. Finally the AI module can make predictions and recommendations. Jimmy Hu told us that more than 50% of newly signed companies will purchase AI modules.
From the data point of view, the customer’s own first-party data value is very important and most effective in employing different marketing scenarios. Third-party data is suitable as a supplement to first-party data.
Many marketing startups of the previous generation around 2010 promoted their own “data banks”, for example, they included a lot of unique third-party data to acquire some customers. But this trend has been declining because the quality and legitimacy of third-party data requires a question mark. The wave of companies that used third-party data for advertising in the past has already encountered bottlenecks. Jimmy Hu’s evaluation of this phenomenon is: “I can only say that Chinese enterprise marketing services may be three to four years slower than foreign markets, but it is interesting that data-driven technology and AI capabilities in marketing scenarios are not lagging behind. This is because we apply technology quickly.”
