一个常见的困惑：中台投入到底值不值？
优品零售是一家拥有三个业态的区域零售集团：120家超市、80家便利店、40家生鲜店。过去几年，每个业态都各自建了自己的系统：
- 超市用的是某品牌ERP，每年license费30万；
- 便利店用的是另一套SaaS系统，每年服务费25万；
- 生鲜店自己找外包做了一套，每年维护费15万；
- 会员系统又是单独买的，每年20万；
- 还有各种小程序、报表工具、数据分析平台……
算下来，每年光软件相关的费用就超过150万。这还不算内部IT团队的人力成本，以及系统之间做对接的费用。
老板一直在犹豫：要不要建个中台，把这些系统都统一了？但一打听，建中台动辄几百万投入，又有点打鼓——这钱花出去，到底能不能赚回来？
这是很多企业决策者的共同困惑。今天我们就以优品零售的案例，来算一笔明白账。
先算清：多套孤立系统的”隐性成本”
很多企业算IT成本的时候，只算license费和服务费。但实际上，多套孤立系统的成本远不止这些。我们把它分成四类：
1. 直接成本：看得见的钱
这是最容易算的，就是每年花在软件上的钱：
- 各系统的license费 / 订阅费
- 各系统的维护费 / 服务费
- 服务器 / 云资源费用（每套系统单独占资源）
优品零售这部分每年大约150万。
2. 对接成本：系统之间”传话”的钱
系统多了，就需要互相打通。A系统的数据要导到B系统，B系统的数据要同步到C系统……这些对接都是要钱的：
- 接口开发费用
- 对接后的维护费用（改一个系统，另一个也要跟着改）
- 数据对不上的时候，人工核对的成本
优品零售每年花在系统对接和数据核对上的钱，大约60万。而且随着系统越来越多，对接成本是指数级增长的——n个系统两两对接，就是n(n-1)/2个接口。
3. 人力成本：养多套团队的钱
每套系统都需要人来维护。ERP要有人懂，SaaS要有人懂，会员系统也要有人懂。IT团队的规模随着系统数量线性增长。
优品零售IT团队有12个人，其中8个人分别负责不同的系统。如果系统统一了，可能只需要4-5个人就能搞定。这部分人力成本，每年大约80万。
4. 机会成本：因为数据不通损失的钱
这是最容易被忽略，但往往也是最大的一块成本。因为系统不通，数据不统一，导致：
- 决策慢：想看个跨业态的报表，要等IT导数据、拼表格，等一周才能出来；
- 营销难：会员数据不统一，同一个人在超市是金卡，在便利店是普通卡，没法做统一的会员营销；
- 库存不准：三个业态的库存各算各的，没法统一调拨，有的店积压、有的店缺货；
- 上新慢：上一个新功能，要在三套系统里各改一遍，周期长、成本高。
这部分成本很难精确量化，但我们保守估计，优品零售每年因为数据不通导致的损失，至少在100万以上。
四笔账加起来，优品零售每年在”多套孤立系统”上的总成本，超过400万。
而大部分企业，只看到了第一笔150万的直接成本，后面三笔隐性成本往往被忽略了。
中台方案：投入多少，省下多少？
我们给优品零售设计的中台方案，核心思路是：建一套统一的业务中台，把会员、商品、订单、库存这些核心能力统一起来，各业态的前台系统（小程序、POS、管理后台）都基于中台来做。
老系统不是一下子全换掉，而是逐步迁移——先迁最痛的部分，再慢慢迁其他的。
投入侧：中台建设要花多少钱？
| 项目 | 费用 | 说明 |
|---|---|---|
| 中台建设（一期） | 180万 | 核心服务：会员、商品、订单、库存 |
| 中台建设（二期） | 120万 | 扩展服务：营销、数据分析、设备管理 |
| 老系统迁移 | 80万 | 逐步把三个业态的系统迁到中台之上 |
| 第一年总投入 | 380万 | |
| 后续年维护费 | 60万/年 | 中台运维 + 迭代 |
看起来第一年投入不小，但我们看看能省多少。
收益侧：能省下多少钱？
直接节省的成本
| 项目 | 改造前 | 改造后 | 年节省 |
|---|---|---|---|
| 软件license/服务费 | 150万 | 40万（保留少量必要系统） | 110万 |
| 系统对接成本 | 60万 | 10万 | 50万 |
| IT人力成本 | 80万 | 40万 | 40万 |
| 小计 | 290万 | 90万 | 200万/年 |
间接带来的收益
这部分更难量化，但影响更大：
- 统一会员营销：三个业态会员打通后，交叉销售带来的销售额提升，保守估计每年增加200万营收（按5%的提升算）；
- 库存优化：统一库存、统一调拨，库存周转率提升10%，减少资金占用约150万；
- 决策效率提升：实时数据、统一报表，管理层决策更快更准，这部分价值很难用钱衡量，但非常重要。
保守估计，间接收益每年在300万以上。
ROI计算
把账算一下：
- 第一年投入：380万
- 第一年节省+收益：200万（直接节省）+ 300万（间接收益）= 500万
- 第一年就回本了，还净赚120万。
从第二年开始：
- 年投入：60万（维护费）
- 年收益：500万+
- 年ROI超过700%。
当然，这是理想情况。实际落地中，间接收益的释放需要时间，可能第一年只能释放30%-50%。但即使按最保守的估计，两年内回本是没问题的。
怎么判断你的企业适不适合建中台？
不是所有企业都适合建中台。我们总结了几个判断标准：
1. 系统数量 ≥ 3套，且还在增加
如果你的企业只有1-2套系统，那建中台可能是过度设计。但如果你已经有3套以上系统，而且每年还在买新系统，那中台的价值就很大了——因为系统越多，对接成本越高，重复建设越多。
2. 有多个业务线 / 业态
如果企业只有一条业务线，一套系统就能搞定，那确实不需要中台。但如果有多个业态、多个品牌、多条产品线，它们之间有共享的核心能力（会员、商品、订单等），那中台就能帮你避免重复建设。
3. 业务变化快，需要快速迭代
如果你的业务很稳定，几年都不用改系统，那中台的价值不大。但如果业务变化快，经常要上新功能、做新活动，那中台的价值就很大——因为中台把通用能力沉淀下来了，前台做新功能就快了。
4. 数据驱动决策的需求强
如果你需要经常做跨系统的数据分析、报表、BI，那数据不通就是很大的痛点。中台统一数据模型之后，数据分析就容易多了。
一个常见的误区：中台不是越大越好
最后说一个常见的误区。很多企业一听中台，就觉得要建一个”大而全”的中台，什么能力都放进去。结果投入巨大，周期很长，最后业务部门还觉得不好用。
我们的建议是：小步快跑，逐步建设。
先从最痛的1-2个领域切入，比如先统一会员，或者先统一商品。把这一块做扎实，让业务部门看到价值，然后再逐步扩展。
优品零售就是这么做的：一期只做了会员、商品、订单、库存四个核心服务，花了180万，6个月上线。上线之后，业务部门马上感受到了好处——会员打通了，商品统一了，上新速度快了。然后二期再做营销、数据分析、设备管理，就顺理成章了。
结语
中台投入到底值不值？答案不是简单的”值”或”不值”，而是要看你的企业处于什么阶段、有多少系统、痛点在哪里。
但有一点是确定的：多套孤立系统的隐性成本，远比你想象的要高。 很多企业只算了看得见的license费，没算看不见的对接成本、人力成本和机会成本。
如果你也在考虑中台建设，不妨先坐下来，把这四笔账好好算一遍。算完之后，你可能会发现，建中台不是”花钱”，而是”省钱”，甚至是”赚钱”。
A Common Dilemma: Is Middle Platform Investment Worth It?
Premium Retail is a regional retail group with three business formats: 120 supermarkets, 80 convenience stores, and 40 fresh food stores. Over the past few years, each format built its own systems:
- Supermarkets use a branded ERP, 300,000 yuan per year in license fees.
- Convenience stores use a different SaaS system, 250,000 yuan per year in service fees.
- Fresh food stores had a custom system built by an outsourcer, 150,000 yuan per year in maintenance.
- The membership system was purchased separately, 200,000 yuan per year.
- Plus various mini programs, reporting tools, data analytics platforms…
Adding it all up, annual software-related expenses alone exceed 1.5 million yuan. And that’s not counting the internal IT team’s labor costs, or the cost of integrating systems with each other.
The boss has been hesitating: should we build a middle platform to unify all these systems? But when he hears that building one costs millions, he hesitates again — will this investment actually pay for itself?
This is a common dilemma for many business decision-makers. Today, let’s do the math using Premium Retail as a case study.
First, Calculate the “Hidden Costs” of Multiple Siloed Systems
When many companies calculate IT costs, they only count license fees and service fees. But the real cost of multiple siloed systems goes far beyond that. We divide it into four categories:
1. Direct Costs: Visible Money
This is the easiest to calculate — annual software spending:
- License / subscription fees for each system
- Maintenance / service fees for each system
- Server / cloud resource costs (each system consumes resources separately)
For Premium Retail, this portion is about 1.5 million yuan per year.
2. Integration Costs: Money Spent on Systems “Talking” to Each Other
More systems means more connections needed. Data from system A needs to go to system B, data from system B needs to sync to system C… all these integrations cost money:
- Interface development costs
- Post-integration maintenance costs (change one system, the other has to change too)
- Manual reconciliation costs when data doesn’t match
Premium Retail spends about 600,000 yuan annually on system integration and data reconciliation. And as systems multiply, integration costs grow exponentially — n systems with pairwise integrations means n(n-1)/2 interfaces.
3. Labor Costs: Money for Maintaining Multiple Teams
Each system needs people who understand it. ERP needs specialists, SaaS needs specialists, the membership system needs specialists. IT team size grows linearly with the number of systems.
Premium Retail’s IT team has 12 people, 8 of whom are responsible for different systems. If systems were unified, maybe only 4-5 people would be needed. This labor cost is about 800,000 yuan per year.
4. Opportunity Costs: Money Lost Because Data Doesn’t Flow
This is the most easily overlooked, but often the largest cost category. Because systems don’t connect and data isn’t unified:
- Slow decisions: Want a cross-format report? Wait for IT to export data and stitch together spreadsheets — a week later you get it.
- Marketing difficulties: Membership data isn’t unified. The same person is Gold tier at the supermarket but Regular at the convenience store — can’t do unified membership marketing.
- Inaccurate inventory: Three formats calculate inventory separately, can’t do unified transfers — some stores have overstock, others have shortages.
- Slow rollouts: Launching a new feature means modifying it in three systems — long cycles, high costs.
This portion is hard to quantify precisely, but we conservatively estimate that Premium Retail loses at least 1 million yuan per year due to disconnected data.
Adding up all four categories, the total annual cost of “multiple siloed systems” for Premium Retail exceeds 4 million yuan.
And most companies only see the first 1.5 million in direct costs — the other three hidden categories are often ignored.
Middle Platform Solution: How Much to Invest, How Much to Save?
Our middle platform solution for Premium Retail has a core idea: build a unified business middle platform that centralizes core capabilities — members, products, orders, inventory — and all front-end systems (mini programs, POS, admin dashboards) across formats build on top of the middle platform.
Old systems aren’t replaced all at once — migration happens gradually: start with the most painful parts, then migrate the rest over time.
Investment Side: How Much Does Building a Middle Platform Cost?
| Item | Cost | Notes |
|---|---|---|
| Middle Platform (Phase 1) | 1.8M yuan | Core services: members, products, orders, inventory |
| Middle Platform (Phase 2) | 1.2M yuan | Extended services: marketing, analytics, device management |
| Legacy system migration | 0.8M yuan | Gradually migrate all three formats onto the platform |
| Year 1 Total Investment | 3.8M yuan | |
| Ongoing annual maintenance | 0.6M yuan/year | Platform operations + iteration |
It looks like a big first-year investment, but let’s see what it saves.
Return Side: How Much Can You Save?
Direct Cost Savings
| Item | Before | After | Annual Savings |
|---|---|---|---|
| Software license/service fees | 1.5M | 0.4M (keep only essential systems) | 1.1M |
| System integration costs | 0.6M | 0.1M | 0.5M |
| IT labor costs | 0.8M | 0.4M | 0.4M |
| Subtotal | 2.9M | 0.9M | 2.0M/year |
Indirect Benefits
This portion is harder to quantify, but the impact is greater:
- Unified membership marketing: After unifying members across three formats, cross-selling revenue increases — conservatively estimated at 2M yuan in additional annual revenue (assuming a 5% lift).
- Inventory optimization: Unified inventory and transfers improve turnover by 10%, reducing capital tie-up by about 1.5M yuan.
- Faster decision-making: Real-time data, unified reports — management makes faster, more accurate decisions. The value is hard to measure in money, but it’s significant.
Conservatively estimated, indirect benefits exceed 3M yuan per year.
ROI Calculation
Let’s do the math:
- Year 1 investment: 3.8M yuan
- Year 1 savings + benefits: 2.0M (direct savings) + 3.0M (indirect benefits) = 5.0M yuan
- Pays for itself in Year 1, with a net gain of 1.2M yuan.
From Year 2 onward:
- Annual investment: 0.6M yuan (maintenance)
- Annual benefits: 5M+ yuan
- Annual ROI exceeds 700%.
Of course, this is the ideal scenario. In practice, indirect benefits take time to materialize — maybe only 30-50% is realized in the first year. But even with the most conservative estimates, payback within two years is achievable.
How to Judge If Your Company Should Build a Middle Platform
Not every company needs a middle platform. We’ve summarized several criteria:
1. System Count ≥ 3, and Growing
If your company only has 1-2 systems, building a middle platform might be over-engineering. But if you already have 3+ systems and keep buying new ones every year, then the value of a middle platform is significant — because more systems means higher integration costs and more duplicate work.
2. Multiple Business Lines / Formats
If your company has only one business line and one system can handle everything, you really don’t need a middle platform. But if you have multiple formats, brands, or product lines that share core capabilities (members, products, orders, etc.), then a middle platform helps you avoid duplicate construction.
3. Fast-Changing Business, Need for Rapid Iteration
If your business is stable and systems don’t need changes for years, the value of a middle platform is limited. But if your business changes fast — constantly launching new features, running new campaigns — the value is huge: the middle platform沉淀 (accumulates) common capabilities, so the front-end can build new features faster.
4. Strong Need for Data-Driven Decisions
If you frequently need cross-system data analysis, reporting, and BI, then disconnected data is a major pain point. Once the middle platform unifies the data model, data analysis becomes much easier.
A Common Misconception: Bigger Isn’t Always Better
Finally, a common misconception. When many companies hear “middle platform,” they think they need to build a “big and complete” one with everything in it. The result: massive investment, long cycles, and business departments still find it hard to use.
Our advice: start small, build incrementally.
Start with the 1-2 most painful areas — maybe unify members first, or unify products first. Get that solid, let business departments see the value, then expand gradually.
That’s exactly how Premium Retail did it: Phase 1 only built four core services — members, products, orders, inventory — costing 1.8M yuan and launching in 6 months. After launch, business departments immediately felt the benefits: unified members, unified products, faster rollouts. Then Phase 2 — marketing, analytics, device management — followed naturally.
Conclusion
Is middle platform investment worth it? The answer isn’t a simple “yes” or “no” — it depends on your company’s stage, how many systems you have, and where your pain points are.
But one thing is certain: the hidden costs of multiple siloed systems are much higher than you think. Many companies only count the visible license fees, ignoring the invisible integration costs, labor costs, and opportunity costs.
If you’re considering building a middle platform, sit down and calculate all four cost categories carefully. After doing the math, you might find that building a middle platform isn’t “spending money” — it’s “saving money,” and even “making money.”