APTnovel Chapter 492 Let the Data Speak
Chapter 492 Let the Data Speak
Tencent acted very quickly, so quickly that even people from JD.com and Ele.me were privately marveling at it.
Liu Chengdong of Ele.me once told Lu Ran on the phone, "You guys have really angered Tencent. They've been going around asking people out for meals lately. Our business line has even received two calls from Tencent, hinting that we should reconsider our cooperation. I ignored them, but their attitude is definitely quite strong."
Lu Ran, holding his phone, leaned back in his office chair. After listening, he didn't show much emotion, simply saying, "I'm not surprised at all that Tencent has the capability."
After hanging up the phone, he sat in his chair and thought quietly for a while.
Tencent certainly has money, manpower, and channels, enabling it to exert its efforts in five or six directions simultaneously.
TUTU doesn't have the financial resources, nor can it make such a large-scale plan.
He can only delve deeply into one or two areas and develop a thorough understanding of them.
Fortunately, depth is sometimes more convincing than breadth.
The data from Ele.me came in earlier than expected.
After the AI delivery dispatch system was integrated with Ele.me, a medium-sized city was selected for a pilot program.
The city has a moderate number of orders and a stable rider team, making it suitable for validating the actual effectiveness of the new system.
After the system ran for a week, the data came out—the average delivery time for riders dropped from 37 minutes to 29 minutes, the empty-run rate decreased by about 20%, and the same number of riders could complete more orders. The delivery cost per order decreased by nearly one yuan.
One yuan may not seem like much, but multiplied by the daily order volume and then by a month, it becomes a considerable sum.
When Liu Chengdong shared this data with Lu Ran on the phone, his tone was noticeably warmer than before: "President Lu, your system is truly effective. I was worried that the data, which looked good in the lab, would be less reliable after deployment, but the data after a week of testing is even more stable than I expected. I'm already having the technical team prepare for the second phase of deployment, planning to expand to two more cities."
After hearing this, Lu Ran didn't seem too excited, simply replying, "Then we'll synchronize it when the second batch of data comes out."
But after hanging up the phone, he walked around the office, his pace a little faster than usual.
JD.com's data came out a few days later, but it was just as good.
After the AI supply chain optimization module was integrated, tests were first conducted on two categories: fresh produce and daily necessities.
These two categories have the highest inventory loss and unsold inventory rates in the e-commerce industry. If you can succeed in these two categories, it will be of greater reference value to other categories.
After running the data for half a month, the out-of-stock rate of fresh produce has decreased by more than 10%, and the unsold rate has also decreased significantly.
The improved inventory turnover efficiency has translated into increased profits, resulting in savings of nearly two million yuan in warehousing and loss costs each month.
Zheng Kai mentioned this figure during an internal report, and it is said that JD.com's senior management is quite satisfied with the result and has approved the full rollout of the AI module to all self-operated product categories.
The data from Kuaishou and Douyin came out last, but it had the greatest impact.
After the AI recommendation optimization module was running on top of Kuaishou and Douyin's recommendation systems for ten days, Zhang Yifan sent over some key data: after integrating TUTU's AI engine, the average user time increased by nearly two minutes.
Two minutes may sound like a small amount, but for a platform like Kuaishou and Douyin, it means a massive increase in user dwell time each day, which translates to more exposure and a higher ceiling for monetization.
In his message, Zhang Yifan used a word that Lu Ran had never seen him use before—"exceeding expectations."
His exact words were: "We made many internal predictions and thought that an increment of one to two minutes was the limit, and it turned out that we were approaching that limit in just ten days. If you continue to optimize, this number might be able to go even higher."
Lu Ran saw the message while he was eating lunch. He put down his chopsticks, read the message twice, and then continued eating. But after finishing the rice in his bowl, he went to get another half bowl.
The data that these three partners produced within half a month was even better than Lu Ran had expected.
It's not that he lacks confidence, but he knows that it takes time for AI technology to be implemented, there are too many variables in real-world scenarios, and the results from the lab often differ from reality.
However, the data generated by the system in the actual scenario was much less biased than he had expected compared to the results he saw in the system.
That's not all the good news.
The three companies' progress has attracted attention in the industry, and Lu Ran is receiving more and more inquiries and exploratory contacts.
The first company to approach us was Tongcheng Travel.
They are among the top-ranked online travel platforms in China with a substantial user base, but their market share has long been suppressed by Ctrip, and they have consistently struggled with traffic and user growth.
People from Tongcheng contacted Chen Mo through Kuaishou and Douyin, saying they wanted to know if TUTU's AI technology could be used for user recommendations and itinerary planning in the tourism industry.
When Chen Mo forwarded this matter to Lu Ran, he added a note: "The other party's attitude is very positive; they must have been interested after seeing the data from Kuaishou."
Immediately afterwards, Didi also sent a message through an intermediary, asking TUTU if it was interested in trying out intelligent dispatching in the travel scenario.
Didi's relationship with Tencent is much deeper than that with Tongcheng. When this intermediary relayed the message, he was very restrained, only saying, "Let's get to know each other first, we may not cooperate."
But Lu Ran knew that given Didi's size and market position, the fact that they would take the initiative to ask meant that they had seen the effectiveness of AI technology in Ele.me's delivery system and were beginning to consider whether they should try it themselves.
Several other small and medium-sized companies have also contacted TUTU – one is a platform for second-hand transactions, one is a tool for online education, and one is an app for local life services.
These companies are not large, but each has a stable user base in its respective niche market.
They approached TUTU for various reasons. Some saw JD.com's supply chain data, while others saw Ele.me's improved delivery efficiency. But the core logic was the same—they wanted to know if they could also achieve such efficiency improvements.
Tencent is also paying attention to these proactive contacts.
According to information obtained by Chen Mo, Tencent has sent people to contact two of the medium-sized companies, proposing resource support and strategic cooperation plans. The intention is clear—to win them over before TUTU can reach an agreement with these companies.
Lu Ran did not rush to take any action after hearing the news.
He sat in his office for a while, sorting out the current situation in his mind.
Tencent continues its long-term strategic layout, acquiring, investing, and cultivating allies, which consumes a lot of funds but has a wide reach.
TUTU focuses on accumulating data, validating value, and attracting new partners in the short term, with low capital expenditure but high depth.
Both sides are going their own way, and whose strategy is more effective depends on a key variable—the time window.
If Tencent's long-term strategy can produce results within a month, then no matter how impressive the data accumulated by TUTU is, it won't be able to stop the enemy from encircling them on a larger battlefield.
However, if Tencent's spending speed and internal coordination difficulties are greater than expected, then TUTU's data accumulation will have enough time to ferment into more cooperation opportunities.
There are only two things Lu Ran can do now.
First, we will closely monitor the data from our three existing partners and continuously optimize the AI engine's adaptability in different scenarios.
Ele.me's delivery system still has room for improvement, JD.com's supply chain module can undergo several more rounds of testing, and Kuaishou and Douyin's recommendation algorithms still have several optimization paths to explore.
These refinements themselves take time, but with each step forward, the data becomes more convincing.
Second, maintain contact with potential partners who proactively approach you, but don't rush to sign agreements.
Let the other party see more data first, and let them judge for themselves whether TUTU's AI technology can really help them.
It would be much easier to discuss specific terms once they themselves have come to the conclusion that it "is indeed useful."
Chen Mo once raised a point during an internal discussion: "While Tencent is making significant moves, they're simultaneously operating on too many fronts. Acquisitions, investments, new businesses—each requires manpower and time to develop. Their energy is scattered, while ours is focused. After a month of this, they might only have one or two lines of substantial progress, while we've already generated quantifiable data across all three lines. If this continues for another month, the balance will continue to tip in our favor."
Lu Ran didn't respond after hearing this, but he agreed with the assessment in his heart.
Tencent and TUTU are completely different paths, and both of them were originally the right ones.
But now, the two paths are about to intersect.
Therefore, identifying the main road is crucial.
Lu Ran smiled.
Let's let the data speak for itself.
...