In 2023, DaaS will enter the "AI Age of Navigation"

In 2023, DaaS will enter the "AI Age of Navigation"

Author|Doudou

Editor|Pi Ye

**Produced by **Industrialist

In 2002, in the highly competitive American Professional Baseball League, the Oakland Athletics were only among the "lower third" in terms of personnel, material equipment, and financial strength.

However, with the help of top students in data analysis, through analyzing data and obscure baseball statistics, the team manager found a group of strange baseball players who had eccentric personalities but had super abilities in baseball.

By breaking through the traditional data management model, we finally achieved impressive results, comparable to those of the powerful New York Yankees.

This is a story adapted from real events and was later made into a movie - "Moneyball". In fact, using data analysis and mining cases to win football games goes far beyond that. The coach of the famous American national basketball team once used the data mining tools provided by IBM to decide on the spot to replace players.

Currently, approximately more than 20 NBA teams use IBM data mining application software to optimize their tactical combinations.

Coaches can use laptop computers to mine data stored on NBA Center servers at home or on the road. Each game event will be classified according to statistics such as points, assists, attacks, and turnovers.

Using big data to win a football match may seem unbelievable, but it has already become an open winning strategy in the industry.

Nowadays, “winning” is also a common proposition for all industries and enterprises. Digitalization has reached the mid-range, data barriers are gradually being broken down, and companies are seeing new ways to grow.

"When we released this DaaS product last year, the amount of investment was basically around hundreds of thousands, two to three hundred thousand, three to four hundred thousand. Starting from Q4 last year, we began to receive millions of investments. Marketing budget." An industry insider said to the industrialist.

DaaS, which has come to the forefront, is entering its era of great navigation.

DaaS, Yunqi

On December 1, 2021, Alibaba Cloud released a new product, DaaS, positioning its core product service as "DaaS". Its essence is to use data-driven growth as the engine to open up and integrate the business flow, data flow and workflow of the enterprise, allowing Data intelligence maximizes value in enterprise production and operations.

It instantly caused a heated discussion in the industry.

In March this year, JD Cloud released the digital intelligence platform "Uplus" for the first time, aiming at brand growth and anchoring the new DaaS track.

The entry of two Internet giants has gradually made this track lively again.

In fact, DaaS is not a new field. In the past few years, the DaaS track has produced several relatively vertical star manufacturers, such as Maicong Software, Youmi Cloud, Huakun Daowei, etc. In addition, some manufacturers involved in data center business have also entered this track long ago.

As early as around 2015, with the advent of autonomous driving, DaaS became very popular.

DaaS (Data as a Service) focuses on providing data from various sources on demand in the form of data APIs. Generally, DaaS platforms also include metadata management, data governance, data development and other functions. Its fundamental role is to help enterprises convert data assets conveniently. Transform it into business capabilities (to respond to the needs for real-time exchange, sharing, and use of data between enterprise applications and systems), and ultimately solve the core growth issues of the enterprise.

At present, in terms of data sources, the DaaS track can be divided into three camps. One is the data-as-a-service platform represented by Alibaba Cloud and JD Cloud, which provide more information based on the second-party data generated by its own retail platform. Extensive data services.

The advantages of Internet giants are: first, because data comes from all aspects of retail, they have greater advantages in marketing. Coupled with the strong ecological construction capabilities and business systems of large manufacturers, they can create advantages from an integrated and full-stack level.

In addition, Internet cloud vendors can also combine their own advantages to empower enterprises and provide some value-added services for DaaS products. For example, JD YouPlus can use its own integrated supply chain advantages to quickly verify its supply chain capabilities when empowering corporate marketing strategies.

The second is vertical manufacturers represented by Datablau Shushu Technology, Shushu Technology, Shence Data, etc. Their data sources come from the business customers' own operations and provide user analysis services for them. The advantage lies in the ability of data cleaning and analysis, and its deep enough penetration in vertical fields that it can achieve a stronger binding relationship with enterprises. For example, Shushu Technology, as a game data analysis service provider, has great advantages in the pan-entertainment field.

The third is manufacturers represented by Youmi Cloud and Tianyancha. The data sources come from public channels and they mainly provide different solutions for different customer groups. The advantage lies in the survey and analysis category, which can meet the data needs of all walks of life.

For example, there is a huge amount of commercial information on products, advertisements, enterprises, consumers, etc. on the Youmi Cloud platform, which allows it to be involved in e-commerce, games, and short dramas, and it has certain advantages.

Generally speaking, these manufacturers each have their own advantages, but they also have some shortcomings.

New players are filing in, old players are continuing to hone their skills, capital is chasing them, and major manufacturers are optimistic. The explosion of the DaaS track is not surprising. But what is not known is that the DaaS industry is not easy to do.

Visible landing is difficult

An e-commerce company has tried to use a DaaS platform to improve its marketing results. The main method is to better understand its customers, provide personalized recommendations and offers, and increase the conversion rate of marketing activities.

However, in practical applications, the company found that the data quality was unreliable, the data collection was incomplete, and the results of data analysis and application were not satisfactory.

First, the company discovered data quality issues when using the DaaS platform, such as missing data, inaccurate data, and outdated data. These issues prevent companies from accurately understanding their customers and making accurate personalized recommendations and offers.

Secondly, the data collection of DaaS platforms is often not comprehensive enough. Although the platform claimed to provide comprehensive data, the company actually found that the platform failed to collect some important data, such as users’ purchase history, browsing history, and search history. The lack of this data makes it impossible for companies to gain a deep understanding of customer behavior and preferences and to conduct precise marketing activities.

Finally, even though the e-commerce company spent a lot of time and resources collecting and analyzing data, the actual results of these data analysis and applications were not ideal. For example, personalized recommendations and offers based on data analysis results did not improve the conversion rate, but instead led to resentment and dissatisfaction among some users.

In fact, DaaS has great application potential in most business scenarios. This is a conclusion that has achieved full consensus. Enterprises are full of hope in using DaaS to improve operations and are actively investing resources in attempts. However, for the vast majority of enterprises, a large number of DaaS projects have not achieved the expected significant improvements.

In summary, problems such as unreliable data quality, incomplete data collection, and poor data analysis and application effects directly reflect the current situation of difficulty in implementing DaaS.

In fact, these problems not only affect the DaaS vendors themselves, but also have a lot to do with their cooperation with ISVs.

For example, when a DaaS platform cooperates with an enterprise, the data provided by the DaaS platform may not be consistent with the data format and standards within the enterprise; there may be multiple different systems and platforms within the enterprise, and the DaaS platform needs to interface with these systems and develop interfaces. and debugging are more difficult; the data processed by the DaaS platform may involve corporate confidentiality and privacy, and effective security measures need to be taken to ensure that the data is not leaked and tampered with, among other issues.

When DaaS cooperates with ISV, the data services provided by the DaaS platform need to be connected to the ISV system through the API interface. However, the API interfaces of different systems are different, and the docking complexity is high, which requires a lot of time and resources for development and debugging.

In addition, the data processed by the DaaS platform is often highly sensitive, such as customer information, transaction data, etc. Therefore, effective security measures need to be taken during the data transmission and storage process to ensure that the data is not leaked or tampered with.

The data business driver behind “True and Fake DaaS”

Yuanqi Forest, a top brand in the industry, has quickly occupied the sparkling water market with the health concept of "0 sugar, 0 fat, and 0 calories". But in the tea drinking industry, competition is extremely fierce, and if you don’t advance, you will retreat. Yuanqi Forest also needs to strengthen its advantages in the sparkling water category and continue to expand the market.

Under this demand, Lingyang customized a DaaS solution for it.

The first is to determine the direction of new product research. Through a hierarchical analysis of the beverage market segment categories, Yuanqi Forest positioned the four major fruit drink trends: sparkling water, tea drinks, plant protein, and fruit drinks. It has been determined that the sparkling water category is still a prominent category in the beverage market, and it can continue its diversified advantages in the sparkling water market and continue to make efforts.

The second is to develop a differentiated marketing strategy. Formulate differentiated marketing directions for new products by describing the characteristics of the four major categories and user insights. For example, the new pineapple flavor of sparkling water is popular according to the season; the new plant protein product is selected to make efforts in the children's market by breaking the circle of high-energy mothers.

Finally, the full-year multi-peak strategy is implemented. 618 launched the first shot of multi-category layout; during the Double 11 promotion period, greater business conversion was achieved, and young people continued to penetrate.

In 2022, Yuanqi Forest’s Double 11 sales increased by nearly 10% year-on-year to 618, and the unit price per customer increased by nearly 10%, and consumers’ willingness to purchase increased. Compared with 618, the asset level of Group A has increased by more than 50%; the growth rate of young people has increased significantly compared with last year’s Double 11, and the brand’s crowd structure has gradually been optimized.

In the case of Yuanqi Forest, we can find several key details for the successful implementation of DaaS.

The first is to determine the direction of marketing data and applications; the second is to continuously gain insights into market changing data with absolute professionalism; the third is to plan a "long-term combat plan" for data management.

This is the key to solving problems and implementing DaaS. However, it is often ignored by many enterprises.

When enterprises conduct data governance, they focus on managing data procedures, scripts, and tasks. This approach prevents the enterprise's data governance from focusing on improving data value. This can result in the accuracy and reliability of the data being compromised, impacting the company's business decisions.

When many enterprises deploy DaaS, their own needs are not met, resulting in misunderstanding-type governance and difficulty in focusing.

In addition, when enterprises conduct data governance, they aim to complete project delivery. However, after the project delivery was completed, the company did not continue to pay attention to the long-term and sustainability of data governance. Therefore, even after project delivery is completed, subsequent data management still lacks continuity and stability.

Management breakpoints make it difficult to unify data, which often leads to a reduction in the data security system, resulting in project-based governance that is difficult to continue.

You must know that the unification of data is the basis for the implementation of DaaS. From this point of view, it does not seem to be a "real DaaS" deployed.

In addition to this, enterprises are managed part-time by employees in data governance. These employees lack professional knowledge and skills in data governance, resulting in unclear responsibilities and low initiative. As a result, enterprises cannot ensure the smooth implementation of data governance. In this case, the quality and reliability of data cannot be guaranteed, and the security and privacy of data cannot be guaranteed.

The enterprise does not have the corresponding IT conditions, resulting in part-time management, which is also an important factor that makes it difficult to implement DaaS. As Cai Ruitao, founding partner & CTO of Youmi Cloud, said, “In today’s digital era, teams that are good at interpreting data will have a huge advantage.”

To sum up, the underlying logic behind the difficulty of DaaS implementation has gradually become clear, namely, "troubleshooting" under misunderstanding management; "true and false DaaS" under project-based management; and "inability to do what one wants" under part-time management.

It is worth noting that the launch of large models may change the dilemma faced by data governance under the traditional model.

Large models make data more valuable

"We have to realize one thing, the brand owner is more professional than us in Know-how. We only have data, analysis capabilities and technology, but he must be more professional than us in Know-how." JD Technology Solutions Zhu Bing, head of the Center’s Growth Solutions Department, told Industrialists.

In fact, as of now, enterprises hold a large amount of valuable data, but the advantages and capabilities brought by this data cannot be replaced by DaaS vendors. As Zhu Bing said, "Overall, we still need to teach our brands autonomy and subjective initiative, and give them to our partners in these categories. I think it is the most professional for him to do this himself."

However, for many businesses, this is a difficult step to take. What is worth looking forward to is that in the era of "big model +", DaaS also has some new possibilities.

For example, in terms of model training and optimization, the DaaS platform can provide a large model training and service platform to help enterprises conduct model training, adjustment, and optimization. Enterprises can use the large amount of data and computing resources on the DaaS platform to train and optimize large models, thereby improving the accuracy and performance of the models.

In terms of model deployment and management, the DaaS platform can provide model deployment and management functions, allowing enterprises to quickly deploy large trained models to the production environment. Enterprises can use the DaaS platform to perform version control and update operations on models to ensure the stability and reliability of the models.

In terms of data preprocessing and enhancement, the DaaS platform can also provide data preprocessing and enhancement functions to help enterprises clean, transform and label raw data for use in training and testing of large models. The DaaS platform can also provide data enhancement functions to improve the generalization performance and adaptability of the model through various transformations and enhancements to the data.

In addition, in the field of natural language processing, DaaS platforms can provide services such as text classification, sentiment analysis, and language generation. In the field of image recognition, the DaaS platform can provide services such as target detection and image segmentation. Enterprises can use these services directly through the DaaS platform without having to build and train models themselves.

Outside of business, security is also a major feature, that is, the DaaS platform can also provide data security and privacy protection functions to ensure the security and privacy of corporate data. The DaaS platform can provide functions such as data backup, recovery and version control to ensure data security and reliability. At the same time, the DaaS platform can also provide functions such as data encryption, access control, and security auditing to protect the privacy and integrity of data.

In other words, under the AI big model, DaaS vendors may accumulate more and more industry know-how, and the needs of enterprises will be increasingly met; the combination of big models and DaaS can enable enterprises to use it more efficiently and accurately. Large model.

Under the traditional model, companies need to build and train models themselves, which requires investing a lot of resources and time. Moreover, the effectiveness of these models may be affected by a variety of factors, such as data quality, algorithm selection, etc. General-purpose large models will also accelerate enterprises to overcome this hurdle.

Overall, "large model + DaaS" will largely change the inherent stubborn problems of current DaaS, leading it to a more benign development and pushing it into the "Age of Discovery."

Perhaps, in the future, every industry can be reshaped by DaaS, just as the movie mentioned at the beginning of the article uses data to "Moneyball".