I started monitoring, but nothing is happening
Whether it's price monitoring or reputation monitoring, some organizations end up in this state after starting and a few weeks pass. The data comes in accurately every day, but no one has acted on that data. They found out that the lowest price had dropped two weeks later, and they first noticed the declining ratings in the quarterly report — they were not monitoring, but just keeping records.
The causes are generally the same. It's because the place where the data accumulates is separate from where people work. The value of monitoring lies not in the amount of data collected, but in how quickly it reaches the responsible person and leads to a response when a change occurs.
Difference between recording and monitoring: Response loop
To make monitoring work, four steps should follow data collection.
1. Defining a baseline
First, determine what is "normal." The normal price range of our product, the usual rating range by store, the usual percentage of negative reviews by channel. Without a baseline, no change can be detected as "abnormal."
2. Detection
Compare regularly collected data with the baseline to identify deviations. These include deviations in the lowest price, appearance of unauthorized sellers, sudden rating drops, significant increases in specific keywords. This step involves selecting only the deviations that meet certain conditions, rather than scanning the table every day.
3. Notification
The detected deviations should reach the responsible person through the channel they work on. It should not only be displayed on the dashboard but also be delivered via email or internal messenger. The response speed of a structure where the responsible person knows "they need to go see" and a structure where they know "they have arrived" is completely different.
4. Response and recording
The responsible person who receives the notification takes action, and the results are recorded back into the data. Whether a warning to an unauthorized seller made them disappear, or responding to negative reviews led to a rating recovery — when the effectiveness of the response is confirmed through data, the loop is complete.
When this loop runs, monitoring changes from being "data compiled at the end of the month" to a "system that remains quiet until needed."
What can be detected
The detection targets for each area are identified as follows.
| Area | Detected Changes | Response |
|---|---|---|
| Price | Deviation in lowest price, non-compliance with recommended price, significant price fluctuations | Channel policy enforcement, promotion response |
| Seller | Appearance of unauthorized sellers, changes in distribution channels | Distribution management, warning measures |
| Reputation | Rating decline, significant increase in negative reviews, specific store issues | Customer service intervention, store improvement |
| Regulation & Compliance | Posts with potential violations, use of prohibited expressions | Review and request corrective action after review |
| Competitors | Registration of new products, changes in pricing policies | Strategic response |
The common point is that it is not about "data accumulating daily" but about "a few deviations from the norm." What the responsible person needs is not a table with 30,000 rows but 7 items to check today.
Designing notifications is half the battle: All notifications are the same as no notification at all
The most common point of failure in the response loop surprisingly lies in the notification stage. There are two ways in which it fails.
One is when there are too many notifications. If every minor change is notified, people will read all of them in the first week, skim through them in the second week, and stop reading them from the third week onwards. Once notifications start to be ignored, truly important changes also get buried.
The other is when there are no notification criteria. Starting with "let's collect everything and put it on the dashboard," it eventually returns to a structure where someone has to go in and check every day.
That's why notifications require design.
- Threshold: Define at what percentage deviation or number of cases to notify. Start conservatively and adjust while operating.
- Grouping and summarizing: Instead of sending individual notifications, group them into 1-2 summaries per day, but separate urgent conditions (e.g., mass unauthorized sales) for immediate delivery.
- Recipient separation: Price notifications go to sales, reputation notifications go to customer service — no one should receive notifications unrelated to their work.
The collection cycle can also be determined from the same perspective. Most responses are sufficient with daily or even hourly regular collection. The important thing is not the frequency of collection but the time it takes from detection to response.
Summary
The purpose of monitoring is not to accumulate data but to respond to changes. The loop of establishing a baseline → detection → notification → response must run for data collection to lead to results.
Hashscraper designs the entire response loop including detection criteria, notifications, and dashboards on top of regular data collection. We take care of crawler operation and maintenance, so clients can focus on making decisions and taking action when notifications arrive.
Recommended Readings
- From web scraping data to decision-making
- Data received in Excel vs. data viewed on a dashboard — how the delivery format changes things
- How to automatically monitor competitor prices using Coupang and Naver web scraping
- How do global brands manage ratings and reviews across dozens of stores in multiple countries at a glance? — Case study on integrated reputation monitoring with Google Maps
Start Now
For the monitoring you are currently running, when was the last time you took action because of the data? We will help you design the collection to notification-response loop together.
