Why do Alibaba’s acquisitions often fail?

The following are excerpts from an article published in Chinese. The author, Dongqi Yu, is a renowned business commentator. The views and perspectives expressed in this article are NOT those of Momentum Works; nonetheless we find them interesting and worth sharing. The current leadership of Alibaba is acutely aware of the problems described here and are taking active steps to fix them. You can also read our book “Seeing the unseen: behind Chinese tech giants’ global venturing” for more case studies, analyses, and reflections on Alibaba and other Chinese tech companies. 

A decade ago, Alibaba was the most formidable company in the internet industry.

From B2B to Taobao, Alipay, Tmall, and Alibaba Cloud.

Alibaba, with its strong drive and organisational capabilities, became a typical example of continuous breakthroughs. Famous for its political commissar system (政委体系), mission vision values, and more, Alibaba has influenced countless entrepreneurs thereafter.

Note: “政委体系” (Zhèngwěi tǐxì) or “political commissar system” at Alibaba refers to a management and organisational structure where certain key individuals or leaders (akin to political commissars) ensure that teams adhere to the company’s culture and strategic goals. These leaders act as internal overseers, aligning various departments with the overarching values and objectives of the organisation.

However, starting in 2019, Alibaba seemed to suddenly find itself in a vulnerable state.

• Alipay faced challenges from WeChat Pay.
• Its desired ventures in content and social media never really took off.
• In food delivery, Ele.me lost to Meituan.
• JD.com, Pinduoduo, and even Douyin’s ecommerce directly invaded Alibaba’s core territory.

In almost every niche market, rivals carved out a piece of Alibaba’s domain.

Why did the once-powerful Alibaba frequently suffer losses in so many battlefields?

Was it because Alibaba’s team wasn’t strong enough?

Obviously not.

For instance, in the local services sector, Meituan’s prized “fast replication ability” actually originated from Alibaba. It was Gan Jiawei (former COO of Meituan and former Sales VP of Alibaba), who was from Alibaba’s “Iron Army” sales team (阿里中供铁军), who brought Alibaba’s management system to Meituan.

Similarly, Didi’s founder Cheng Wei (程维), and many successful entrepreneurs and executives, have Alibaba backgrounds.

Today’s Alibaba still has the most mature group of managers among all companies.

Also, Alibaba has a comprehensive training system and a thorough methodology about business and management. Inside Alibaba, there are countless methodologies, taught to every new employee and manager through training.

So, what is the real reason behind Alibaba’s losses in multiple key battlefields?

To find out, I communicated extensively with many people who worked at Alibaba and its competitors and identified the core logic behind Alibaba’s problems.

My final judgement is –

For Alibaba, which once excelled in organisational capabilities, today’s problem ironically stems from the organisation itself.

Next, I’ll start with one of Alibaba’s noticeable problems.

 

1. Why do Alibaba’s acquisitions often end in failure?

It must be said that Alibaba is still the most successful company in China when it comes to inspiring employee morale.

Its mission, vision, values, political commissar system, and incentive-based management keep Alibaba’s employees highly motivated. This is something Alibaba is well aware of.

So, whenever Alibaba acquires a company, it always brings its organisational capabilities into play. In most cases, the work atmosphere within these acquired companies quickly becomes vibrant. However, you’ll notice that many of Alibaba’s acquisitions, such as Yahoo China, Wandoujia (​​豌豆荚), UC Browser (UC浏览器), Youku (优酷), Xiami Music (虾米音乐), Koubei (口碑网), Ele.me, etc (饿了么)., have seen more failures than successes.

On Maimai 脉脉 (a professional social networking platform), there’s a pattern described for Alibaba’s acquisitions: from acquisition and Alibaba staff entering, rapid loss of existing employees, to the eventual shutdown of the business. 

Both the loss of original staff and the shutdown of business are obviously not what Alibaba initially intended. 

Alibaba wanted to retain the existing employees because just having the brand and users isn’t enough to sustain the business without people familiar with it. However, regardless of the intention, once acquired by Alibaba, the original staff tend to leave.

Usually, the story of a company after being acquired by Alibaba goes like this:

First, during the acquisition, Alibaba’s political commissars bring in their management system.

Alibaba is known for “heavy assessments and incentives.”

Internally, Alibaba has always adopted a 361 assessment method.

30% of the people are “exceeding expectations,” receiving the best performance reviews and extra incentives, making it easier for them to get promoted.

60% are “meeting expectations,” which is average.

Additionally, unless the overall team performance is exceptionally good, there will definitely be 10% of people being rated as “below expectations.” Not only do they miss out on promotions and year-end bonuses, but being rated “below expectations” twice in a row could also lead to termination.

The 361 assessment injects vigour into Alibaba’s team, keeping their spirits high.

When political commissars enter an acquired company, they replicate Alibaba’s management actions. Indeed, under these actions, the employees’ work status improves significantly. However, many businesses don’t fully align with Alibaba’s management system.

For example, when switching from the original management system to the political commissar system, employees’ ranks and salaries are realigned. But aligning two different pay systems is not so easy. Many original employees, feeling their ranks and total income have dropped, become dissatisfied and leave. This happened with many of Ele.me’s old business executives. 

The key point is, for a food delivery service like Ele.me, the platform can only deliver food. Cooking still depends on the restaurants. Service to customers requires a large number of these merchants, who are maintained by business executives.

Many of these business executives left Ele.me with their merchant resources and went to Meituan, causing a significant impact on Ele.me’s business.

Second, old Alibaba employees, in large numbers, enter the new business.

The entrepreneurs whose companies were acquired usually received money and achieved financial freedom, and they wouldn’t want to play second fiddle. When the founding team leaves, Alibaba sends in its own CEOs, CFOs, and other core executives.

In addition to core executives, many Alibaba employees also transfer to these acquired companies.

In Alibaba, transfers are allowed, even encouraged, for those below P9 level. Alibaba has an organisational department that decides the roles of employees at P10 level and above, the top talent. But employees below P9, as long as they have completed 2 years in their old role and 1 year after a promotion, and their performance isn’t too poor, can transfer.

Moreover, when transferring, as long as the new business accepts them, the old business can’t retain them.

The new Alibaba managers in the acquired business need familiar people to work with, so a large number of old Alibaba employees come to the new business.

When Ele.me was acquired, many former “Taobao customer service reps” and sales personnel from Alibaba’s “Iron Army” sales team, and operations personnel from the Taobao system, also followed to Ele.me.

Third, the old Alibaba employees in the new business, unfamiliar with it, still start working immediately.

People go to the new business not just because of interest but more because of opportunities.

After all, it’s easier to achieve success in a new business with a lower starting point and a budget than in a mature, established business.

Thus, the newcomers have expectations of quick results and rapid promotions.

Regardless of their familiarity, they will start working right away.

Fourth, newcomers, less knowledgeable than the old employees, often end up receiving more resources.

Logically, resources are limited, and since Alibaba people just entered the new business and don’t understand it, resources should be allocated first to the old employees who know the business.

However, it turns out that newcomers often get more resources.

The reason lies in:

On one hand, new managers parachuted in by Alibaba also want quick results but don’t understand the business. When the knowledgeable veterans explain the business, they don’t grasp Alibaba’s jargon. Newcomers may not understand the business, but neither do their managers, yet they are more versed in Alibaba’s language and know how to “tell the story.”

Thus, newcomers (from other Alibaba business) can communicate with managers more effectively than veterans from the newly acquired business.

On the other hand, many Alibaba people come in groups. A P9 leader (director) brings along many P7s (managers) and P8s (senior managers) who have followed them for years. They naturally want to give these P7s and P8s opportunities to achieve results, reflecting the so-called “clan spirit.”

As a result, newcomers end up getting more resources.

Fifth, shouldn’t this issue with resource allocation be corrected when it’s time to evaluate results? Often, it isn’t.

When it comes to evaluations, newcomers, not understanding the business and performing poorly, shouldn’t receive good ratings. Veterans, who understand the business, should be rated better, right? 

In reality, it’s often the opposite: newcomers who don’t understand the business receive better evaluations, while the hardworking veterans from the acquired business lines often get lower ratings.

This is quite puzzling. How do these poorly performing projects receive good evaluations? It’s mainly done by focusing on “narrative” and “long-term impact.”

For example, after being acquired by Alibaba, Ele.me began having two “operational campaigns” every winter and summer. These were held not due to real peaks in user demand but because there are two evaluation periods in a year, necessitating two major events to boost performance.

For instance, during one campaign, a business person in charge of 5 “hives” (delivery zones) had targets for both GMV growth and surpassing Meituan in 3 hives, both high-weighted targets. But increasing GMV was too difficult, so they focused on the target of “3 hives surpassing Meituan” instead. They concentrated their budget on the smallest 3 hives, making it easy to achieve the “surpassing Meituan” effect. The moment the budget was concentrated, the outcome was certain. 

Overall, their GMV growth wasn’t significant, but the target of surpassing Meituan in 3 hives was beautifully achieved, resulting in a good evaluation. However, this approach had no practical significance for the business. Concentrating subsidies on smaller hives is unsustainable; once subsidies are withdrawn, things revert to their original state.

However, just meeting targets isn’t enough; it also requires managerial approval to impact evaluations.

Then comes the question.

Can’t the managers see the issues behind these targets?

Of course, they can.

But if you were a manager also under the pressure of a 361 assessment, wanting the best performance, would you prefer to point out issues, showing your subordinates did poorly, or turn a blind eye, making the performance look better? Obviously, the latter, as it’s more beneficial for the manager’s performance and promotion.

And when it’s possible to take extra care of one’s own people, it’s done, reflecting the “clan spirit.”

Newcomers who can spin narratives well are at an advantage during end-of-year performance evaluations compared to veterans unfamiliar with these rules. As a result, newcomers often get higher evaluations, while the diligent veterans receive lower ratings. Feeling unfairly treated, veterans further leave.

Sixth, meanwhile, newcomers would launch many “innovation” projects.

Under the incentive and pressure of the 361 assessment, everyone aims to be in the top 30%.

How to be part of this 30%?

Purely relying on diligently doing business might not significantly outperform others’ data.

To become part of the top 30% and achieve an “exceeding expectations” rating, innovation is key.

On one hand, innovation is more likely to “stand out.”

On the other hand, established businesses have high data requirements, but “new stories” have lower initial expectations, making them easier to achieve.

“Starting new projects is much simpler than optimising existing business!”

It’s not only business personnel who think of “innovation.”

To make end-of-year reports look better, various resources also gravitate towards these “new stories.”

Launching innovations without truly understanding the business is unrealistic from the start. However, when newcomers know they can manipulate narratives and hide certain long-term impacts, and when managers often collaborate, making these “innovative” projects “seem successful” becomes much easier.

In projects where user number is key, it is easy to boost user number with subsidies,  in the last few months before performance review, even if these users won’t retain.

For sales-driven projects, over-promising can inflate sales figures, even if these customers are unlikely to repurchase the following year.

Moreover, established projects have some set of metrics, but new projects often don’t.

In Alibaba, many new projects follow the “shoot first, draw the target later” approach (先开枪再画靶).

Once the story is convincing, they start working, and only at the end of the year do they consider which metrics to apply. This allows for more operational flexibility.

By year-end, through new projects, newcomers often get better performance reviews, even promotions.

However, the project that was highlighted in the previous year often fizzles out in the next.

Because these new projects hide many pitfalls.

Continuing them the next year often means “fixing / filling in the pits before making incremental progress. (先填坑、再做增量)” 

It’s far easier to completely abandon old projects and start a new story afresh.

Thus, by the second year, these once highly praised projects have failed. But the people involved were promoted in the previous year.

Often, after promotion, these people are transferred to other areas.

Seventh, these Alibaba employees don’t stay long in new ventures.

They often move to the next seemingly more opportunistic new business after two or three years, post-promotion.

Ultimately, resource wastage due to failed innovations, many decisions made by those “still unfamiliar with the business,” and employees focused on “storytelling” rather than doing good business, lead to efficiency loss in operations.

Unscientific evaluations drive away stable, experienced veterans.

Personnel instability leads to each generation having new ideas, lacking policy continuity, and being more prone to storytelling.

Business performance declines year by year, with many ventures ending in failure.

Yet, among these shuttered business lines, many managers who successfully climbed the ranks were born.

 

2. Why do these issues leading to the failure of acquired businesses occur at Alibaba?

This “acquisition to failure” story is somewhat understandable in its first half, as it could happen in most companies.

After all, aligning management systems is essential for smooth operations.

As long as talent mobility is allowed, people will naturally gravitate towards more promising opportunities.

Everyone wants to work and achieve results quickly.

These are common phenomena in large companies.

However, the latter half is abnormal:

People who don’t understand the business can get resources, unsuccessful projects can receive high ratings, anyone can innovate whenever they want, and failed projects can be abandoned easily.

All these seem unlike what would happen in a mature company with established systems and methodologies.

For instance, the failure of “Rural Taobao” (农村淘宝) is a complete manifestation of the lack of standards and methods in project initiation, model design, and management:

In 2016, Alibaba aimed to develop Rural Taobao with two goals: one, to genuinely do charity work and improve shopping for rural residents; two, to attract rural users to Taobao.

How to make ecommerce shopping available to rural residents? The main barrier wasn’t the products, but logistics.

So, Rural Taobao introduced a new model.

First, a “Rural Taobao” app was created, where orders placed would be sent to a warehouse in the county town.

Then, a “Village Partner” was recruited for each area, similar to today’s community group buying leaders, who would transport packages from the county town back to the villages and distribute them to residents.

Up to this point, it seemed like a good “story.”

However, in 2018, Alibaba realised a problem:

Rural Taobao was intended to drive new traffic to Taobao, but why create an independent app that diverts traffic?

The independent app had to be shut down.

So, all users were directed to Taobao, and “Rural Taobao” became a tab within Taobao. This change led to a significant loss of “Village Partners.”

Because, in the first two years of promoting “Rural Taobao”, salespeople from the “Alibaba’s Iron Army” sales team, transferred to Rural Taobao, promised the Village Partners: As long as you’re a “Village Partner,” you’ll get a commission from all orders placed in your area through Rural Taobao for life.

Relying on this promise, the salespeople beautifully achieved their targets, leading to many of them getting promoted.

However, in 2018, when Rural Taobao was no longer an independent app, users could order both Rural Taobao products and other items from Taobao’s main platform. The income of Village Partners dropped sharply, leading to their massive exodus, and the original model of Rural Taobao became unsustainable.

Yet, these problems could have been avoided:

First, there should be standard methods for new project initiation.

At least the direction, model, and competitiveness should be clearly addressed.

If there were standards, the creation of an independent app, which contradicted the project’s original goal, could have been avoided at the design stage.

Alibaba’s goal with Rural Taobao was traffic, so how could the ecommerce benefits from this traffic be given to Village Partners for life?

Such a distribution method should have been ruled out during the design phase.

Second, there should also be standard methods for monitoring and management.

The excessive promises made by salespeople and the high profits for Village Partners, given Alibaba’s expertise in managing numerous sales teams and operations, should have had standard methods established from the beginning for management and monitoring.

But none of these were done.

So, in 2018, the Village Partner model collapsed dramatically.

Later, the team that took over Rural Taobao unsurprisingly abandoned the previous model.

Today’s Rural Taobao has transformed into offline stores in county towns selling home appliances.

And such a lack of standards and methods is reflected in almost every business problem at Alibaba: 

First, there’s no standard or method for resource allocation.

Within Alibaba, many teams can start working on projects that interest them if they have spare capacity. If they need help from other divisions, as long as they’ve been at Alibaba for a while, they can gather some familiar colleagues to start.

When they need resources, they seek their managers, adding resources layer by layer.

However, some projects have great opportunities and fierce future competition. Escalating them layer by layer with insufficient resources could mean missing opportunities.

Some projects are high-risk and hard to validate. Giving out resources without strong strategic abilities could lead to resource wastage.

These should be problems solved by standards and methods, but at Alibaba, it mainly depends on “storytelling.”

Second, there’s no standard or method for replicating results during operations.

For example, in Hema (Alibaba’s grocery business), a certain category in the Shanghai store sells exceptionally well.

Managers want to extract and spread the store manager’s experience, but the manager says, “I want to enjoy the benefits a bit longer, help me hide this for 3 months! After 3 months, we can talk about it!”

Based on “clan spirit,” managers would really help them hide it.

Effective actions that should be quickly replicated and promoted nationwide end up being limited to a single region.

Third, the phenomenon of “shooting first, drawing the target later” indicates there’s no method for setting goals.

It’s for this reason many projects become a vacuum with no goals or assessment standards.

The problems caused by the lack of standards and methods result in “too low a floor” for actions.

What does this mean?

Normal work should generate a net positive for every action. Only those whose benefits outweigh resource consumption should receive positive evaluations.

But at Alibaba, it’s not the case—driving away veteran employees, new businesses with no follow-up, hiding expertise and refusing to replicate—all these harm the business, causing losses far greater than the value created.

If there were standards and established work methods, these behaviors should not happen. Yet, not only do they occur, but they also help the perpetrators get high evaluations.

Doesn’t Alibaba know the importance of standards and methods?

Of course, Alibaba does!

Alibaba is one of the companies that values methodological accumulation the most. It has developed many methodologies and standard management methods, taught to all employees through mature training. However, these methodologies often seem to disappear in practice. 

First, every new employee at Alibaba undergoes weeks of off-the-job training, including not only value indoctrination but also business experience sharing.

Second, everyone promoted to P8, becoming a manager, undergoes a management training called “Xiake Xing” (侠客行), receiving almost a full set of management tools.

Setting goals first is a standard management action taught in the training. However, this standard action disappears in many innovation projects, hence the emergence of “shooting first, drawing the target later.”

Additionally, Alibaba has a famous saying: “A process without results is nonsense, and results without a process are rubbish” (没有结果的过程是狗屁，没有过程的结果是垃圾). 

Gan Jiawei, a former core member of Alibaba’s Central Supply Iron Army, also famously said, “Let 80% of people achieve 80% of the sales champion’s performance” (让80%的人，做到销售冠军的80%好). 

These are the bases for extracting business experience and replicating it quickly. 

However, the ability to replicate quickly disappeared in many departments at Alibaba later on.

Third, regarding how new projects are conducted and initiated, Alibaba also has a standard methodology called “One Map, One Heart, One Battle,” which, if followed, assures “certain victory” (“一张图、一颗心、一场仗”，只要做到，就“一定赢”). 

This methodology dictates that every initiative must start with a “unified strategic vision.” After establishing this vision, everyone should align with “one heart” in their thinking. Finally, tactical planning is completed for “one battle.” However, during the initiation of new projects (like Rural Taobao, which clearly violated the strategic consensus by creating an independent app and failed to control over-promises tactically), this methodology also seems to disappear.

This is the real reason behind Alibaba’s difficulties:

It’s not the lack of methodologies or an attempt to establish standards. Rather, just having accumulated knowledge and training does not necessarily lead to effective standards.

The real issue lies in Alibaba’s lack of a stable “referee.”

 

3. Merely discussing methodologies is futile. Only with a stable referee can standards and methods become actions for everyone.

Compared to Alibaba, which places immense emphasis on methodologies and has a heavy training system, Meituan, with an even stronger “methodology atmosphere,” doesn’t have as long training periods. However, actions like “fast replication” and “benchmarking everything” have become almost universal behaviour at Meituan.

The difference lies in the “referee.”

Whether methodologies are formed and whether everyone does the right thing doesn’t depend on training but on the quality of the referees. Only when referees block shortcuts will the correct actions occur.

I asked a friend who switched from Alibaba to Meituan, “Do you still have to ‘tell stories’ at Meituan?”

He said, “Storytelling is useless at Meituan; you can only work hard.”

Where does the difference lie between the referees at Meituan and Alibaba?

For example, Taobao is Alibaba’s most mature business, while Ele.me has always been learning and benchmarking against Meituan.

Taobao and Ele.me are already the departments with the most comprehensive data within Alibaba.

At Ele.me, assessments mainly focus on “data results.”

How are the food delivery salespeople assessed?

There are two main metrics:

The first is “new volume,” looking at the number of new merchants, how many days in a month these new merchants operate for more than 8 hours, and whether they have more than 10 orders. If so, these merchants are considered to have survived.

The second is “operations,” mainly looking at GMV. Also, how many merchants maintained by this salesperson sign up when Ele.me organises events.

As long as these two sets of data results are achieved, one would definitely get a decent evaluation at Ele.me.

As for how these were achieved, no one cares, and even if they wanted to, they couldn’t see.

However, at Meituan, a business would have four to five hundred metrics, covering not just new stores but also how many dishes they list, whether they participate in basic activities, and whether they offer discounts or other promotions—all reflected in the data.

These hundreds of metrics are not only related to the final business results but also to all potential short and long-term impacts and all processes.

Moreover, in this set of metrics, besides the result metrics, all process-related metrics are also part of the assessment for the business executives.

So, if you think, “I’ve driven up the GMV, so I should get a good performance review, right?”

That’s not enough; these process metrics must also be achieved.

These process metrics, like listing menus, configuring discounts, etc., essentially describe the standard operational actions of a store.

As long as these are completed, the store’s performance won’t be too bad.

As a result, the motivation for business executives to falsify data is greatly reduced.

However, what if business executives still want to manipulate the system? For instance, the behaviour at Ele.me where they create “boost” in business performance by concentrating budget in a specific area.

Such actions are practically impossible at Meituan.

At Ele.me, each business person has a discretionary subsidy amount they can freely allocate to any area or store they choose, just by setting it up in the system.

But at Meituan, you’ll find that you can’t just set it up like that. Every store in the system has a predefined subsidy cap, say, a maximum of 2 yuan per order. You can’t exceed it, so there’s no room for business executives to tweak the budget.

However, with these metrics in place, surely someone has to monitor them, right?

Just having metrics is useless if managers don’t pay attention. At Meituan, people are specifically assigned to do this.

Each business area has hundreds of metrics, further broken down by objectives and cities. A single business can easily generate tens of thousands of different data metrics.

What’s frightening is that these tens of thousands of metrics are reviewed weekly:

The business analysis team reviews these tens of thousands of metrics every week, identifies all anomalies, and compiles them into an email to the business managers.

Each week, they compare data to the previous week, month, and the same period last year, ensuring any potential issues are discovered promptly.

Discovered issues are communicated to relevant business personnel in advance for solution proposals.

Whether these solutions are actually implemented is followed up by a PMO (Project Manager), and only after the business analysis team verifies that the anomalies have disappeared do they stop reporting them.

For stable businesses with consistent metrics, this process of running numbers and comparisons can be automated. But for businesses with unstable metrics, manual intervention by the business analysis team is required. No problem – they’ll handle it until the business stabilises.

What about storytelling for new activities?

There are always methods that no one has tried before and for which the business analysis team lacks experiential data.

For example:

Once, a business person at Ele.me found that subsidising the lowest-priced merchants was more effective for boosting orders. The logic is simple: subsidising 10 yuan on a 100 yuan meal is less noticeable and less attractive. But subsidising 10 yuan on a 20 yuan spicy hotpot is much more tempting! So, subsidising the lowest-priced merchants seemed like a great story, enhancing the perceived value of the same subsidy and appearing as a significant advancement in subsidy logic.

But the problem behind this is that these users are not retainable.

The large number of users attracted by the lowest price are either highly price-sensitive, switching to Meituan tomorrow if it offers better subsidies, or simply can’t afford takeout regularly.

In reality, what’s often hidden in these stories is the long-term impact of such activities. But at Meituan, your story might not have even been fully told or reached evaluation time before the business analysis team discovers “an issue with user retention this month.”

The users brought in, who were supposed to repurchase within 3 days, didn’t. This information would be discovered during the weekly data review and reported to managers, naturally preventing the story from being completed.

Later, Ele.me continued to subsidise the lowest-priced merchants for a long time. In contrast, Meituan quickly realised that users attracted to low-priced merchants were not valuable and stopped supporting business personnel in subsidising these merchants. This restriction was also embedded in the system, eliminating the chance for business personnel to try the same story again.

In all these “stories”, the easiest aspect to hide is the long-term impact:

Subsidising low-priced users without repurchase; overpromising sales, which hurts retention; giving too many orders to a leading merchant, which might lengthen pickup times and harm user experience.

These are the main areas of storytelling.

It’s also where referees find it hardest to see clearly.

To clearly understand each action’s long-term impact and calculate the complete cost-benefit ratio to ensure it truly leads to “overall optimisation”, Meituan’s business analysis team spends a lot of effort studying the mathematical relationships behind all long-term impacts.

The results of these studies are compiled into documents that every business analysis team member can refer to when evaluating related issues.

Thus, long-term impacts can’t be hidden from the business analysis team.

Storytelling that trades long-term harm for short-term numbers doesn’t work at Meituan.

That leaves only one path.

New projects!

Old projects are so tough, but are new projects easier in terms of metrics?

After all, “starting new projects is much simpler than optimising existing business at Alibaba”!

But this is not normal!

The reason it’s simpler at Alibaba is that you can “shoot first, draw the target later.”

Under normal circumstances, how could successful innovation be easy?

Successful innovation is much more difficult than continuous optimization!

To ensure that each innovation doesn’t blindly waste resources, at Meituan, initiating an innovative project is quite challenging. To obtain resources, one must withstand rigorous scrutiny and clearly answer questions about “user value, industry chain, model, and resource endowment.” To answer these clearly, the initiator needs considerable strategic capability.

Additionally, as soon as the business starts, the business analysis team, acting as referees, enters the scene. Concurrently with the business operation, the metrics for judgment are rapidly formed—these metrics keep increasing each month and stabilise after a few months.

When Meituan established such a strict and stable referee system, changes occurred:

First, every manager naturally focuses on and manages every action of their employees. Normally, who would want to delve into such minutiae? It’s much easier to just look at results. But not managing actions is unacceptable; if you don’t supervise your subordinates’ actions, your process metrics can quickly become problematic. Even if the process metrics are fine, if the actions are inadequate, the business analysis team won’t let you off the hook on long-term metrics. What to do? Only truly effective practices that are good for merchants and customers can pass on all metrics.

Second, fast replication becomes the shortest path to business growth. True successful innovation and finding new breakthroughs are incredibly difficult. As the business matures, it only gets harder. Thus, benchmarking others and replicating quickly become the fastest, most cost-effective, and most certain methods to improve performance. Under the pressure of the referee system, fast replication becomes Meituan’s organisational capability.

Third, even without training, methodologies rapidly spread. Just like replicating business actions, without the need for training indoctrination, everyone will naturally seek effective methods. Effective methods quickly spread and become consensus within the company.

Ultimately, the quality of the referees determines whether “athletes” follow the rules of the game. Meituan’s data-based, strictly enforced referee system blocks most shortcuts.

However, at Alibaba, such an evaluation system is not as robust as Meituan’s.

Alibaba leans more towards “rule by individuals.” In most cases, the referee is not data but people. Although Alibaba is also an internet company that looks at data for evaluations:

First, the granularity isn’t sufficient to accurately cover increasingly complex businesses and fully understand actions and comprehensive long-term impacts.

Second, Alibaba only has a generally standard data approach but often lacks a unified business model to evaluate business efficiency. Alibaba’s data analysts manage reports and maintain data standards, but rarely do they independently complete evaluations.

Third, sometimes the necessary data is missing, and “athletes are nearing the finish line, but the referee hasn’t even arrived yet”.

When many positions at Alibaba can select partial data to present in PPTs during promotions and performance reviews, it creates space for people to adjust narratives and self-promote, making the quality of evaluations highly dependent on the individual evaluators.

However, as long as evaluations are left to “people,” loopholes are almost inevitable.

On one hand, people’s evaluations are unstable; not every manager may fully understand the business or have the ability to assess accurately.

On the other hand, every manager’s interests align more closely with their subordinates. Intentionally or unintentionally, they may “go easy” during evaluations.

At the root, all employees, whether frontline staff or managers, have interests that don’t fully align with the company’s. Their primary concern is how to get promoted, increase their salary, and obtain more stock options.

Only the boss’s interests align completely with those of the company.

Therefore, in areas unseen by the boss, actions that harm the company for personal gain inevitably occur.

In such a large company with so many people, how can the boss see everything?

The solution lies in establishing a referee system that represents the company’s interests to help the boss “see.”

When such a stable, independent referee system isn’t established, countless “loopholes” appear in management.

Each loophole becomes a “shortcut” for employees:

• After all, if the referee can’t see, it’s easier to tell stories and cobble together metrics than to do good business.
• After all, compared to further improving already high metrics, it’s much easier to tell a story of going from 0 to 1.
• The supposed hardest task of “going from 0 to 1″ at Alibaba paradoxically seems easier—”innovating is simpler than improving data.” This isn’t describing a business reality but a shortcut for “innovation projects, regardless of quality, are easier for telling stories and getting good evaluations.”

“Clan spirit” is the source of loopholes under rule by individuals.

“Storytelling” is the method of taking shortcuts within these loopholes.

More alarmingly, “taking shortcuts” becomes a case of “bad money driving out good”. When storytellers achieve better evaluations and faster promotions, those competing with them are also forced to learn to tell stories. As storytellers rise to positions of greater power, they become more accustomed to “going easy” and teaching these shortcuts to their subordinates.

Consequently, when Dai Shan 戴珊 managed Taobao’s overseas operations, she had to form a new reserve management team composed of P5-P7 level employees, creating a small organisation in the hopes that they could penetrate the existing layers of management and bring real issues to light.

Thus, due to the lack of a robust referee system, Alibaba is plagued with problems where shortcuts are ubiquitous:

• In acquired businesses, “attention” brings opportunity, leading old Alibaba staff to swarm in, eventually causing the original staff to leave and the business to fail.
• Internally, with numerous people innovating and many new projects dying within a year, the people responsible can still get promoted.
• With shortcuts available, why bother doing business properly? Actions that harm the organisation keep occurring.

In fact, this is not unique to Alibaba. For any company, establishing a high-quality referee mechanism to block shortcuts is essential for proper conduct to prevail. If you aspire to establish standards and popularise methods, they must rely on a strict referee system. Only when every employee must rely on standards and methods to achieve sufficiently good business results can the widespread adoption of methodologies and the implementation of standards become a reality.

 

4. How did Alibaba gradually reach its current state?

Today’s problems at Alibaba stem from its past success. From 2000-2010, Alibaba was the pioneer of ecommerce. Back then, its challenge wasn’t beating competitors but convincing the market filled with merchants who didn’t understand or believe in online sales.

Alibaba was the market’s educator, needing to get more people to understand and accept ecommerce.

Being a market educator is always tough. To break through, employees needed strong beliefs, robust execution, and a continuous stream of innovative ideas. This became the starting point for Alibaba’s organisational capabilities:

First, to foster strong beliefs, Alibaba created a culture of sharing success stories and internal PR, so everyone could see the progress of their peers, fostering a winning mindset.

Second, since the road of a first mover / market educator is tough, Alibaba’s teams were naturally encouraged to bond closely, fostering a culture of camaraderie, sincerity, and mutual support.

Third, to drive strong execution, Alibaba implemented a 361 culture. Forced ranking meant that regardless of overall team performance, there had to be a distribution of 361—except in cases of exceptional team performance. High incentives and generous benefits were offered. To receive these, it wasn’t enough to “do well”; one had to “do better than others.”

Forth, a culture of strong push came from the need for robust execution. For important tasks and key battles, pressure was exerted at every level to ensure faster and better completion.

Fifth, encouraging and accommodating innovation was necessary to find new paths for business. Any idea was worth trying. Allowing and even encouraging employee transfers to pursue personal interests was seen as beneficial for both employee development and innovation. 

After all, innovation can’t come from top-down pressure but must arise from genuine “belief” and “willingness.” At Alibaba, not only were failures in innovation tolerated, but every innovative project was also more likely to receive resources. Any employee with the energy and network could gather resources for innovation—if everyone was willing, they could start right away. Every manager was also allowed to allocate resources to innovative projects.

This more “sales-driven” culture was once Alibaba’s mainstream. The infamous “new employee ice-breaking session” primarily occurred during this period. Although Alibaba’s main business shifted from sales-driven 1688 (wholesale / B2B marketplace) to product and operation-driven Taobao and Alipay after 2015, the culture from the sales system still largely influenced Alibaba’s flavor.

From the positive side: a lively atmosphere and high incentives greatly boosted Alibaba’s morale and fighting spirit. Among BAT (Baidu, Alibaba, Tencent) companies, Alibaba was known for its hard working culture. Allowing bottom-up innovation led to many new opportunities, such as:

• Yu’E Bao 余额宝 (cash management offering within Alipay) , which popularised financial management among white-collar workers, wasn’t a top-down strategic product but originated from the Alibaba team’s bottom-up design. Yu’E Bao gave many users their first reason to download the Alipay app, leading to the later success of Ant Financial.
• The birth of DingTalk (钉钉) came from Chen Hang 陈航 (nickname: Wu Zhao 无招), the former CEO of Alibaba DingTalk, trying to find a way forward for his team after the failure of Alibaba’s previous attempt at its social app, Laiwang (来往). The team secretly developed DingTalk over several months. Today, DingTalk has a daily active user base of 200 million.

However, like every organisational form, Alibaba’s approach had its downsides. Seeds of problems were always present. Since the overall evaluation was still “rule by individuals,” it was manageable for old core businesses—at least the company had a sense of how to assess and evaluate. But for new businesses that Alibaba “didn’t understand” and the “innovation” that everyone wanted and could do, storytelling as a shortcut and data packaging as a shortcut were always present.

Someone will always discover shortcuts first. Those who discover and exploit them taste success first. Once one person benefits, more people learn and use these shortcuts, and when they get promoted, they influence others, expanding the impact of shortcuts.

In a culture of “rule by individuals,” the origin of loopholes is “clan spirit,” while “storytelling” becomes the method within these loopholes.

The systems that helped Alibaba succeed amplified the power of these shortcuts:

• Under the pressure of the 361 system, what should have been limited competition to achieve goals turned into an endless race of “if you’re good, I must be better.” Merely achieving goals isn’t enough for a good evaluation because there’s always someone innovating or creating highlights. In reports, there must be breakthroughs and innovations; after all, “a bad report means a year wasted” at Alibaba. To get a 3 (high rating), you must innovate and create “highlights” just like others.
• Internal PR also affects evaluations. Normally, meeting targets doesn’t merit PR. To get noticed in internal PR, one must find a way to create “highlights”.
• Alibaba’s internal innovation was rampant but lacked the capability to evaluate innovation and didn’t have enough qualified innovative talent. Thus, creating highlights relied more on “storytelling” and year-end PPTs.
• High-pressure assessments and a strong push culture may work in sales teams, where performance can be objectively measured. But in product and operations, what should be a year-long project, when scrutinized and intensified by higher-ups, becomes a 3-month deadline. Under such pressure, actions inevitably become distorted, and when these distorted actions yield poor results, the only option is to resort to reporting and storytelling.
• The value of “loyalty and righteousness” can easily morph into a “you scratch my back, I scratch yours” clan mentality.
• Ultimately, storytelling and a culture of universal innovation lead to significant resource waste.
• When storytelling suffices for good evaluations, fewer people bother doing real business.
• Confident in its organisational capabilities, Alibaba’s political commissar system brought such a situation to every new business. As soon as a commissar arrives, 361 and battle reports are implemented, but regardless of whether the business needs innovation or, like Ele.me and Hema, more grounded efforts in delivery and quality control.

In summary, Alibaba’s management system unleashed tremendous energy among its employees. But in the absence of proper referees, much of this energy was misdirected, leading to greater wastage the more energy there was.

Once, there was a force that could guide Alibaba’s energy in the right direction, acting as a sort of referee. This was Alibaba’s strong “value culture.” But compared to data-based refereeing, value culture is less precise and stable. Values depend on belief, which can be established through continuous success but may not withstand adversity. Post-2019, with the setbacks faced by Ant Financial, defeats in core ecommerce battles against JD.com and Pinduoduo, and events challenging its values, Alibaba’s “value-based refereeing” became unprecedentedly fragile.

In fact, a stable, data-based refereeing system is the kind of referee that can function consistently, regardless of circumstances.

Moreover, times have changed. It’s not that bottom-up innovation desired by Alibaba is no longer valuable, but the era is no longer just about innovation and finding ways to educate the market. Today, ecommerce has become widespread. As new opportunities on internet platforms gradually diminish, companies like JD.com and Pinduoduo are directly challenging Alibaba’s territory, while Meituan and ByteDance enter a stalemate competition with Alibaba. The theme of business has changed. 

The new theme is not only “innovation” but also “efficiency.” 

The competition on “selection, speed, quality and savings” (多快好省) ultimately boils down to efficiency.

Resources that could have been used to reduce prices for customers or provide better service are wasted on each failed innovation attempt.

 

5. What should Alibaba do next?

In an era of innovation, where capturing new opportunities was crucial, the wastage of resources was a secondary concern. However, as the era of efficiency competition arrives, many previously insignificant issues have become critically important:

• For instance, the effectiveness of innovation—whether each innovation can succeed and how much value it can bring—has become important. It’s not about “not innovating,” but rather that more thorough validation is needed before innovating.
• For example, balancing short-term and long-term impacts—In the early stages of business, problems are generally more apparent, and the ROI of most actions is easier to calculate. But as the most significant issues are addressed, more and more problems will revolve around this balance, such as, “Promotions can boost GMV, but they also challenge logistics capacity, leading to longer fulfilment times. Longer fulfilment times may hurt customer retention. So, is this action worthwhile? Is the increase in GMV more important than the impact on retention?” Such questions will become increasingly numerous and critical.

For every Alibaba business, establishing a data-based refereeing system and rules for resource allocation has become unprecedentedly important. And accomplishing this is more challenging than one might expect.

Calculating short-term impacts is relatively straightforward, like understanding GMV and subsidies. But long-term impacts—like the interplay between merchant benefits and retention, or the long-term impact of customer experience on brand value—are more complex.

For Meituan, the large business analysis team of thousands of members is not just about numbers. They also put significant effort into researching long-term impacts, trying to understand the interplay between short-term and long-term effects. Unsurprisingly, even many of Meituan’s own staff may not fully grasp the value of this work, often seeing business analysis as engaging in seemingly unnecessary research.

So, when business analysis can act as a referee for business decisions, who then referees the business analysis team? Who can accurately judge whether their research is truly valuable, accurate, and reliable?

This requires managers to truly understand and use the data, to unravel the logic behind complex variables. The ones who can judge the business analysis referee are typically the top management of the company—often the highest executive officers.

In conclusion, for Alibaba to navigate this new era effectively, it’s not just about unleashing employee energy or fostering innovation. It’s about channelling these energy and innovation efforts more efficiently and effectively, ensuring that every initiative contributes meaningfully to the company’s long-term success. This approach requires a shift from a predominantly people-driven evaluation system to a more data-driven, systematic approach, where decisions and evaluations are based on comprehensive, long-term analyses rather than short-term gains or narratives.

Therefore, the construction of such business analysis systems at Meituan and Kuaishou is no accident—their leaders are all from science and engineering backgrounds. They believe in data, understand the value of data, and are convinced that data can effectively model the real world. This belief is why they set such high standards for their business analysis teams and why they consistently invest in them.

Wang Xing and other senior executives at Meituan are serious about using data and are extremely sensitive to it. Wang Xing remembers specific data points from a report presented half an hour earlier. Data is his most familiar tool and the “lens” through which he views the world.

However, in an era where “data” can achieve an unprecedentedly accurate depiction and evaluation of business, such data-refereeing capabilities have already become essential for any company aspiring to achieve true efficiency.

For Alibaba, building this new organisational capability is the only way forward. They must either find executives with a “data eye,” fully empowering them with the goal of building a stable refereeing system, or their top decision-makers must learn to view the entire business world through data, just like Meituan.

 

6. Conclusion

Alibaba was once the internet company that most effectively fostered employee initiative. This led to strong execution and innovation capabilities, enabling Alibaba to educate the market during the ecommerce era. However, it also led to various problems Alibaba faces today.

At its core, the issue within Alibaba is the lack of a stable, evaluative referee system. When the top management lacks the tools to see through “stories,” and middle and lower management, in their own interest, join employees in storytelling, the absence of a sound evaluation system creates loopholes. These loopholes, in turn, influence employee behaviour through the omnipresence of shortcuts. Consequently, Alibaba’s acquisitions often end in failure, and many internal initiatives for new businesses hardly survive past their first year. In the face of shortcuts’ allure, earnest work seems increasingly less cost-effective.

The fundamental issue is not that Alibaba’s once-prized innovation capacity has weakened, but that the times have changed. What’s needed now is not just innovation, but achieving higher operational efficiency than competitors. To form an organisation capable of efficient operation, Alibaba needs to establish a data-based evaluation system.

But can Alibaba build such a system?

Currently, I see two unfavourable factors:

Firstly, in other companies, the development of this capability comes from leaders who understand and believe in data. At Alibaba, however, Jack Ma has been out of the picture for quite some time. Even if he were present, this likely wouldn’t be his forte. To Alibaba employees, Ma always seemed more like a “teacher”—genuinely wishing the best for his employees, providing them with training, and allowing them the freedom to choose their future directions and undertake their desired projects. Old Alibaba staff might see “Teacher Ma” as a great humanist and a good person in the liberal arts. But such a liberal artist might not possess the data-driven perspective needed to see the world.

At the moment, Alibaba’s top management consists of professional managers. Can professional managers, tasked with entrepreneurial responsibilities, lacking the influence and interests fully aligned with the company that come with ownership and equity, truly and effectively drive this change? I don’t have the answer.

Secondly, the sheer diversity and dispersion of Alibaba’s businesses today far exceed those of other companies. Unlike Meituan, which primarily deals in low-margin businesses, or ByteDance, which focuses mainly on content, today’s Alibaba operates in 2B and 2C markets, both domestically and internationally, online and offline, making its business incredibly complex. Each of these varied businesses requires its own distinct organisational capabilities.

Alibaba has chosen to empower the leaders of each business unit, granting them the authority to determine their own local organisational strategies. Alibaba’s partners need to compete in their respective fields against the best entrepreneurs like Wang Xing (Meituan) and Colin Huang (Pinduoduo).

However, Alibaba’s partners each have their own methodologies, unlike Meituan, where “executives are cut from the same cloth.” This diversity in approaches among Alibaba’s partners leads to inconsistency in business outcomes:

• Hema, which should have been able to rapidly replicate successful operational results, failed to inherit Alibaba’s sales system’s replication capabilities.
• Ele.me, which should focus on data, data-driven decisions, and meticulous cost management, ended up overly pursuing “short term performance / highlights” and “innovation” after the entry of political commissars.

No business leader on the battlefield likes to be interfered with. Thus, Alibaba’s decentralised structure makes building a unified refereeing system even more challenging, facing opposition and challenges from each independent business unit.

Building such a refereeing system becomes even more difficult.

However, changes are quietly happening within Alibaba: 

Ele.me has always had the strongest data-driven culture within Alibaba, though this data mechanism was not established by Alibaba but during the era of Zhang Xuhao (张旭豪), founder of Ele.me, before its acquisition. This data system, compared to other Alibaba businesses, is somewhat reminiscent of Meituan.

In 2021, Yu Yongfu (俞永福) took charge of Ele.me. Under his leadership, many reports shifted from story-heavy PPTs to logic-focused Word documents. Internal resource allocation also saw standardisation for the first time: new projects must answer several specific questions. The most important metrics during assessments shifted from GMV to ROI. Ele.me is increasingly resembling Meituan.

With staff reductions and efficiency improvements, new opportunities are no longer as tempting or easy to quickly succeed in. Consequently, more Alibaba employees are hesitant to transfer to new positions, as failing in a new role could lead to higher risks of being laid off, compared to staying in their current positions where risks are lower. For the first time, internal personnel movement within Alibaba has slowed down.

With the disappearance of the shortcut of changing roles, more and more people are starting to seriously consider “how to succeed in their current business?” 

After all, the targets hanging over their heads might now be a long-term commitment, rather than just a two-year stint before moving to a new business line.