Chapter 171 New Business

Cao Shen told Cheng Zhuo about the plan to do food delivery and set up a "delivery" joint venture, and Cheng Zhuo's eyes turned green.

Seeing the popularity of food delivery, he had wanted to do it for a long time, and several core members of the company had been sharpening their knives for a long time.

Every time he saw a food deliveryman from another company passing by, his eyes turned red and hot.

They also did in-depth research for a long time.

But I was embarrassed to mention this to Mr. Cao. First, it was Mr. Cao who pointed out the way for us to do the cash register system and "Tongfubao" offline agent.

It is indeed because of following Mr. Cao's instructions and relying on Mr. Cao's core resources that the company has achieved today's results.

Now that I see the popularity of food delivery, I want to transform. I feel like I have "deviated from my original intention".

Secondly, other companies are now competing for various payment entrances, and food delivery is an important high-frequency payment usage scenario, but Mr. Cao is as stable as a rock.

No matter how I think about it, I feel that something is wrong with this matter.

So I thought that Mr. Cao, who is a technical person, is probably not interested in these "physical businesses".

Every time I think of this, the courage I originally had to discuss it is lost.

This cannot be blamed on Cheng Zhuo and others for being timid, the key is that the boss is Cao Shen.

Has Mr. Cao made any wrong decisions since he started his business?

No! Absolutely not!

Not only has he not, but the path he has pointed out every time is extremely correct.

Just like this matter at home, if Mr. Cao had not pointed out the direction at the beginning, the company would have collapsed long ago.

How could it be like now, not only with income and profit, but also with several rounds of financing, becoming a hot commodity in the capital market.

Under Mr. Cao's undefeated "record", anyone is prone to self-doubt.

If Mr. Cao doesn't do something, there must be something wrong with it.

So, Cheng Zhuo and others also repeatedly pondered what was wrong with the takeaway business, and the research was very in-depth.

As a result, they really found the problem!

Influenced by Cao Shen, Cheng Zhuo and his team all became technology geeks.

When they were researching takeout, they were thinking about how to maximize the efficiency of "physical work".

After much research, they felt that the key was the "best path" for picking up and delivering meals.

In this case, we can "empower" through technical means!

The best path, write an algorithm to calculate it!

The idea is very good, but after a long time of research, they found that it was the development of this intelligent system that stumped them.

This system, in the final analysis, is an "intelligent scheduling system".

But it contains a lot of modules. Only by improving and researching technology from prediction, mining, dynamic pricing, planning, scheduling, intelligent operation and hardware can the real business be implemented.

And in this process, a balance must be achieved between experience, efficiency and cost.

If Cao Shen knew that Cheng Zhuo and his team were doing this technical research, he would definitely tell them to "give up first".

Because artificial intelligence technology is the basis of this "intelligent scheduling system".

And in 2012, artificial intelligence obviously had not developed to this level.

Moreover, artificial intelligence requires a large amount of data for training and optimization.

Cheng Zhuo and his team haven't started food delivery yet, so where can they get the data to "feed" the intelligent algorithm?

At this stage in the original world, all food delivery is "stupid delivery".

Of course, by around 2018, this set of things has been relatively available in the original world.

Many intelligent algorithms have been modularized and even open source.

To build this system, you must first have an algorithm platform to perform large-scale real-time feature calculations and have the ability to machine learning.

Second, your scheduling system must be able to achieve multi-point and multi-person global optimization, which requires ultra-large-scale real-time computing.

Third, you also need to develop a regional planning system to dynamically divide the distribution area and plan the scope of merchants.

Fourth, there must also be an IOT Internet of Things system to grasp the status and location of riders and vehicles in real time through smart electric vehicles and wearable devices on riders.

Fifth, the ETA estimated arrival time system must have accurate predictions and deep learning models for all links.

Accurate predictions are made for the expected delivery time, the merchant's meal delivery time, and the future order load of the business district, so that the order can be matched with the best path under the optimal decision to ensure the maximum efficiency of order acceptance, the best time for picking up and distributing meals.

Sixth, LBS is based on the geographic location service system, including the 5-level address library for delivery, rider route optimization, etc.

Seventh, the perception system, the merchant's geographic location fence and motion status recognition.

The so-called geographic fence technology means that when you enter a specific area, the mobile phone can automatically obtain relevant information about the area, and once you leave, it will be gone.

This is one of the important supporting technologies for dispatch optimization.

Cheng Zhuo and his team are also stubborn. They are doing their own research when there is not enough "technical preparation" in this world.

It's probably equivalent to junior high school students doing math problems for postgraduate entrance examinations.

I can only say that I admire their spirit of research emotionally.

As a result, of course, they didn't get it, after all, the research time was not long.

They found that it was barely okay with just a small amount of data, but it would be a disaster if there was too much data, and the complexity would increase exponentially.

Finally, I found that you have to come up with several sets of "intelligent algorithms" for this thing.

Moreover, I have never figured out what type of algorithm can achieve the best efficiency.

Earlier, they had studied “graph data”, but later found that it was a huge project.

Moreover, this system is not good enough for testing. If you want to run real data, a large amount of data, the algorithm may fail.

It is not smart at all, and it is not as fast as the deliveryman’s own judgment based on experience.

In addition, a large amount of data also involves various real-time and high-concurrency processing issues. It is really impossible to handle it.

Of course, this is too absolute. If you spend enough time and enough cost, you can get it sooner or later.

But at that time, the daylily was cold.

They seemed to have figured it out at that time:

The reason why Mr. Cao did not do takeout is because “smart delivery” is too complicated?

If it is like other companies, there is no smart order-taking and delivery system, “stupid delivery”, piling up people and competing for manpower. They carefully made models and found that it was too cruel to burn money like this.

In this case, you must continue to raise funds to survive, which is really a game of “fighting to the end” and “survival of the fittest”.

The risk is too high! After all, the competition in the food delivery business is so fierce now, and there are "giants" behind each company.

So that's why Mr. Cao must have carefully calculated not to do this.

It's really high!

After thinking it through, I was very unwilling to not be able to do food delivery.

Once, there was a market with great potential worth hundreds of billions in front of me, but I didn't have the skills.

If I were given another chance, alas, I would have no choice.

I can't come up with an intelligent delivery system!