一个现实的困境：“想数字化，但设备太老了”
邻邻便利是一家有300多家门店的区域连锁便利店品牌。他们想做数字化升级，但遇到了一个很现实的问题：门店里的冷柜、冰柜、饮料展示柜，大部分都是用了五六年的老设备，根本没有联网功能。
如果全部换成智能冷柜，一台就要七八千，300家店每家5台冷柜，算下来就是上千万。对一个区域连锁来说，这个投入太大了。
但不换吧，问题又很多：
- 冷柜温度全靠店长每天手写记录，漏记、瞎记是常态，食安风险大；
- 饮料卖完了不知道，经常出现”顾客想买、冰柜空了”的情况，损失销量；
- 冷柜耗电不知道，有些门店电费异常高，查不出原因。
有没有办法不换设备，也能让老设备”变智能”？
答案是：有。用”外挂式”改造方案。
核心思路：外挂式改造，不碰设备本身
老设备智能化改造，很多人第一反应是”换控制板”或者”改电路”。但这样做有两个问题：一是需要专业人员上门，人工成本高；二是改了电路之后，设备出了问题厂家不保修。
我们的方案走了另一条路：完全不碰设备本身，所有改造都是”外挂”的。
就像给手机贴个钢化膜、装个手机壳——手机本身没动，但功能增强了。
具体来说，方案分四层：
┌──────────────────────────────────────┐
│ 应用层：小程序 │
│ 店长巡检 / 总部监控 / 告警推送 │
├──────────────────────────────────────┤
│ 平台层：业务中台 │
│ 数据存储 / 分析 / 告警 / 报表 │
├──────────────────────────────────────┤
│ 网关层：边缘网关 │
│ 数据采集 / 本地缓存 / 协议转换 │
├──────────────────────────────────────┤
│ 感知层：外挂传感器 │
│ 温度 / 电流 / 门磁 / 人体红外 │
└──────────────────────────────────────┘下面逐层说。
感知层：外挂传感器，贴上去就行
最底层是传感器。这些传感器全部是免安装、外挂式的：
- 温度传感器：贴在冷柜内壁，用3M胶一粘就行，测柜内温度；
- 电流传感器：卡在冷柜的电源线上（钳形传感器，不用剪线），测设备运行电流，判断压缩机是否在工作、有没有异常耗电；
- 门磁传感器：贴在冷柜门上，测门开了多久、开关次数；
- 人体红外传感器：贴在冷柜上方，测有没有人在冷柜前停留，结合销售数据可以分析”看了但没买”的转化率。
所有传感器都是电池供电，续航1-2年，不用布线，不用接电。
安装有多简单？店长自己看个视频就能装，一台冷柜5分钟搞定。这意味着300家店不用派工程师上门，省了一大笔人工成本。
网关层：边缘网关，一个店一个
传感器的数据通过蓝牙或者LoRa传到边缘网关。一个门店放一个网关，就能接店里所有的传感器。
网关的作用不只是”传数据”，它还做了几件很重要的事：
1. 本地缓存，断网不丢数据
门店网络不稳定是常态。网关本地能存30天的数据，断网了继续存，网络恢复了自动补传。中台侧做幂等处理，不会产生重复数据。
2. 边缘计算，减少上行流量
不是所有数据都需要传到云端。比如温度数据，网关本地先做判断：温度正常就不上报，只在温度异常或者每小时汇总的时候上报一次。这样能大大减少流量消耗，尤其是在4G网络的场景下。
3. 协议转换，统一格式
不同传感器用的协议不一样（蓝牙、LoRa、433MHz等），网关把它们统一转换成MQTT协议，再发到中台。这样中台只需要处理一种协议。
网关本身成本也不高，几百块钱一个，比换一台冷柜便宜多了。
平台层：业务中台，数据的归宿
数据到了中台之后，就可以做各种事情了：
1. 实时监控
总部大屏上能看到所有门店、所有冷柜的实时温度、电流、门开关状态。哪台冷柜温度异常，一眼就能看到。
2. 智能告警
温度超标、电流异常、门没关好……这些情况中台会自动告警，通过小程序推送给店长和区域经理。
以前冷柜坏了，可能要等第二天店长上班才发现，里面的饮料全化了。现在温度一异常，马上就能收到通知，几小时内就能处理，损失大大减少。
3. 数据分析
中台积累了数据之后，可以做很多分析：
- 耗电分析：哪些门店、哪些冷柜耗电异常？是不是该除霜了？是不是老化了？
- 补货分析：结合门开关次数和电流数据（压缩机工作频率间接反映饮料量），可以预判什么时候需要补货；
- 食安合规：自动生成温度记录报表，满足监管要求，不用店长手写了。
应用层：小程序，人人都会用
最后，数据怎么呈现给人用？答案是小程序。
我们做了三个端的小程序：
- 店长端：每天打开小程序就能看到店里设备状态，有没有告警，需不需要处理。巡检的时候扫码确认，自动留痕；
- 区域经理端：看自己管的几十家店的整体状态，哪些店有问题，直接安排处理；
- 总部端：看全公司的数据大屏，各种统计报表。
为什么用小程序而不是App？因为不用安装，不用培训，店长扫码就能用。对于一线员工来说，学习成本越低，用起来的概率越高。
投入产出：花小钱办大事
算一笔账。邻邻便利300家店，每家5台冷柜：
| 项目 | 单价 | 数量 | 总价 |
|---|---|---|---|
| 温度传感器 | 50元 | 1500台 | 7.5万 |
| 电流传感器 | 80元 | 1500台 | 12万 |
| 门磁传感器 | 30元 | 1500台 | 4.5万 |
| 边缘网关 | 500元 | 300店 | 15万 |
| 中台+小程序 | - | 1套 | 20万 |
| 合计 | 约59万 |
不到60万，就能把1500台老冷柜全部”智能化”。如果换智能冷柜，同样的数量要上千万。
效果呢？上线半年后：
- 冷柜故障导致的商品损耗，从每月约8万降到了每月2万，一年省72万；
- 电费优化（发现异常耗电的冷柜及时维修/更换），单店月均电费降了约5%，一年省约30万；
- 食安巡检人工成本，从每月3万降到了每月5千，一年省30万。
一年就能回本，之后每年都是净收益。
写在最后
传统设备智能化改造，不一定要”推倒重来”。很多时候，老设备还能用，只是缺了”眼睛”和”耳朵”。用外挂传感器+边缘网关+小程序+中台的方案，花很少的钱，就能让老设备接入数字化体系，产生实实在在的价值。
数字化不是只有”买新的”这一条路。找到合适的切入点，用最小的投入撬动最大的价值，才是中小企业数字化的正确姿势。
A Real Dilemma: “We Want Digital, But Our Equipment Is Too Old”
NeighborMart is a regional convenience store chain with 300+ locations. They wanted to upgrade digitally but hit a very real problem: most of the refrigerators, freezers, and beverage display cases in their stores were 5-6 years old with zero networking capability.
Replacing them all with smart refrigerators would cost 7-8 thousand yuan per unit. With 5 refrigerators per store across 300 stores, that’s tens of millions — way too much for a regional chain.
But without replacing them, the problems pile up:
- Refrigerator temperatures rely entirely on handwritten logs by store managers, with missed entries and fake records being the norm — major food safety risk.
- When beverages sell out, nobody knows. Customers want to buy but the fridge is empty — lost sales.
- Nobody tracks refrigerator power consumption. Some stores have abnormally high electricity bills but can’t figure out why.
Is there a way to make old equipment “smart” without replacing it?
The answer: yes. With a “bolt-on” retrofit approach.
Core Idea: Bolt-On Retrofit — Don’t Touch the Equipment Itself
When people think about smart retrofitting for old equipment, the first thought is often “replace the control board” or “modify the circuits.” But this approach has two problems: first, it requires professional technicians on-site, which is labor-intensive; second, modifying circuits voids the manufacturer warranty.
Our approach takes a different path: we don’t touch the equipment at all — all modifications are “bolted on.”
It’s like putting a screen protector and case on your phone — the phone itself doesn’t change, but its capabilities are enhanced.
Specifically, the solution has four layers:
┌──────────────────────────────────────┐
│ Application: Mini Program │
│ Manager Inspection / HQ Monitor / │
│ Alert Push │
├──────────────────────────────────────┤
│ Platform: Business Middle │
│ Data Storage / Analytics / Alerts │
│ / Reports │
├──────────────────────────────────────┤
│ Gateway: Edge Gateway │
│ Data Collection / Local Cache / │
│ Protocol Conversion │
├──────────────────────────────────────┤
│ Sensing: External Sensors │
│ Temperature / Current / Door / PIR │
└──────────────────────────────────────┘Let’s go through each layer.
Sensing Layer: External Sensors — Just Stick Them On
The bottom layer is sensors. All sensors are installation-free, bolt-on:
- Temperature sensors: Stick to the inside of the refrigerator with 3M adhesive, measures internal temperature.
- Current sensors: Clamp onto the refrigerator power cord (clamp-style sensor, no wire cutting needed), measures operating current to determine if the compressor is working and if there’s abnormal power draw.
- Door contact sensors: Stick on the refrigerator door, tracks how long the door is open and how many times it opens.
- PIR motion sensors: Mount above the refrigerator, detects if someone is standing in front of it — combined with sales data, you can analyze “looked but didn’t buy” conversion rates.
All sensors are battery-powered with 1-2 year battery life — no wiring, no power connection needed.
How easy is installation? Store managers can do it themselves after watching a video — 5 minutes per refrigerator. That means no need to send engineers to 300 stores, saving a fortune in labor costs.
Gateway Layer: Edge Gateway — One Per Store
Sensor data transmits via Bluetooth or LoRa to the edge gateway. One gateway per store connects all sensors in that location.
The gateway does more than just “pass data along” — it handles several important functions:
1. Local Cache: No Data Loss When Offline
Unstable store networks are the norm. The gateway stores up to 30 days of data locally. When the network goes down, it keeps storing; when the network comes back, it auto-resumes upload. The middle platform uses idempotent processing to avoid duplicates.
2. Edge Computing: Reduce Upstream Traffic
Not all data needs to go to the cloud. For example, with temperature data, the gateway makes local judgments: if temperature is normal, don’t report — only report when temperature is abnormal or in hourly summaries. This drastically reduces data traffic, especially on 4G networks.
3. Protocol Conversion: Unified Format
Different sensors use different protocols (Bluetooth, LoRa, 433MHz, etc.). The gateway converts them all to MQTT before sending to the middle platform. This way, the platform only needs to handle one protocol.
The gateway itself is also low-cost — a few hundred yuan each, much cheaper than replacing a refrigerator.
Platform Layer: Business Middle Platform — Where Data Goes
Once data reaches the middle platform, you can do all kinds of things with it:
1. Real-Time Monitoring
HQ dashboards show real-time temperature, current, and door status for every refrigerator across all stores. Any refrigerator with abnormal temperature is immediately visible.
2. Smart Alerts
Temperature excursions, current anomalies, doors left open — the platform automatically alerts, pushing notifications to store managers and regional managers via mini program.
Previously, if a refrigerator broke, you might not discover it until the next day when the manager came in — by then, all the beverages inside had thawed. Now, the moment temperature goes abnormal, you get a notification and can handle it within hours — losses are drastically reduced.
3. Data Analytics
Once the platform accumulates data, you can do many analyses:
- Power consumption analysis: Which stores, which refrigerators have abnormal power draw? Time for defrosting? Showing signs of aging?
- Restocking analysis: Combining door open counts and current data (compressor frequency indirectly reflects beverage levels), you can predict when restocking is needed.
- Food safety compliance: Automatically generate temperature log reports to meet regulatory requirements — no more handwritten logs by store managers.
Application Layer: Mini Program — Everyone Can Use It
Finally, how do people interact with the data? The answer: mini program.
We built three mini program interfaces:
- Store Manager端: Open the mini program daily to see store equipment status, any alerts, what needs attention. Inspections are confirmed by scanning, automatically recorded.
- Regional Manager端: See overall status of the dozens of stores you manage, which stores have issues, assign and track resolutions directly.
- HQ端: Company-wide dashboards and various statistical reports.
Why mini program instead of an app? Because no installation, no training — store managers scan a QR code and start using it. For frontline employees, the lower the learning cost, the higher the adoption rate.
ROI: Small Investment, Big Returns
Let’s do the math. NeighborMart has 300 stores, 5 refrigerators each:
| Item | Unit Price | Quantity | Total |
|---|---|---|---|
| Temperature sensors | 50 yuan | 1,500 units | 75,000 |
| Current sensors | 80 yuan | 1,500 units | 120,000 |
| Door contact sensors | 30 yuan | 1,500 units | 45,000 |
| Edge gateways | 500 yuan | 300 stores | 150,000 |
| Middle platform + mini program | - | 1 set | 200,000 |
| Total | ~590,000 yuan |
Under 600,000 yuan to “smarten up” 1,500 old refrigerators. Replacing them with smart units would cost tens of millions for the same quantity.
And the results? Six months after launch:
- Product loss from refrigerator failures dropped from ~80,000 yuan/month to 20,000 yuan/month — saving 720,000 yuan per year.
- Electricity optimization (finding and fixing/replacing abnormally consuming refrigerators) reduced average per-store monthly power bills by about 5% — saving ~300,000 yuan per year.
- Food safety inspection labor costs dropped from 30,000 yuan/month to 5,000 yuan/month — saving 300,000 yuan per year.
Pays for itself in one year. After that, it’s all net gain.
Closing Thoughts
Smart retrofitting of legacy equipment doesn’t have to mean “tear it all down and start over.” Often, old equipment still works fine — it just lacks “eyes and ears.” With external sensors + edge gateways + mini program + middle platform, you can bring old devices into the digital system for very little money and generate real value.
Digital transformation doesn’t have only one path — “buy new ones.” Finding the right entry point, using the smallest investment to unlock the most value — that’s the right approach for SMB digital transformation.