Defining your data collection strategy for WeChat touch points

There are three possible levels of analysis: WeChat native analytics, WeChat API level analytics and how we can go beyond.

Getting the data that you need to run your WeChat Mini Program (WeChat MP) is difficult: understanding your customer, segmenting your audience and iterating over the data you collect is virtually impossible with the native capabilities of the WeChat MP SDK.

We won’t go into detail about the WeChat reporting interface, as Chelsea, from our marketing team has previously covered that. Read the full translation of the UI in English as a starter to understand what the platform capabilities are.

There are three possible levels of analysis:

  • WeChat native analytics
  • WeChat API level analytics
  • How to go beyond

WeChat Routine Analytics

Routine Analytics are Tencent’s term for the daily metrics you would use for high level understanding of your mini program.

Good questions to ask yourself before diving more into the data collection are: what are my needs, what’s my budget, and how mature is my organization regarding data analysis?

Collecting data is great but making sense out of it is better. If you’re just looking for simple information such as base growth of your followers, most popular articles, menu clicks… then the native reporting is perfectly suited for your needs.

All of that data is available natively, without any particular setup. You won’t have to implement anything on top of the initial SDK.

You can already get some interesting information with native capabilities. I’d pay particular attention to:

  • bounce rates (an all time favorite, regardless of the analytics scenario)
  • traffic acquisition by QRcode - tag your QRcode to see how much traffic from individual sources (useful for calculating ROI and the best place to focus resources)
  • retention - in a highly competitive market, knowing how successful you are in bringing your audience back is critical

Other native metrics are pretty much vanity - for example, user base growth, which is useful only for popularity contests, and not in any kind of business case.


Not everyone will want (or need) to prototype, iterate and optimize conversion rates. However, this is all lagging data that won’t give you any space to explore, build hypotheses, experiments and predictions. You will understand how your mini program is performing (to some extent) but not why.

WeChat Customized Analytics

Custom reports will let you implement events that pass values to a dataLayer and conversion funnels.

With some front-end resources, budget, and enough time to collect and analyze your data, you should find some interest in customized analytics features.

Typical events you will want to configure on an e-commerce platform:

  • Product views
  • Add to cart
  • Orders in process
  • Transaction success

The data fields work as a dataLayer, allowing you to pass values with each event. In the screenshot above, for each .addToCart events, you will pass values defined by data variables.

Another interesting feature, even though it’s hard to scale, is funnel setup.

Just like conversion funnels by Google Analytics, you will be able to understand drop-off and conversion from one step to another along your sales pipeline. For simple setups, funnels mix both events and pageviews for a convenient view of your flow.

Customized analytics gives you enough information to get conversion averages for your users. You will be able to understand what your total audience understands of your design choice and thus, improve your mini program UI to push users to wherever you want to drive them.


To be fair, custom analytics will allow you to implement a simple build, measure, learn data loop - by using events to question your design choices.

However, it doesn’t solve two critical issues for me:

Audience segmentation

Any analysis how-to will tell you that from the first page that data doesn’t answer anything when it’s not sliced into segments.

You need to understand the experience of your key segments and use the data collected to engage, reengage and convert. Analyzing averages won’t lead you anywhere.

The WeChat reporting UI won’t let you import your customer segments and won’t give you the tools necessary to build complex segments (e.g. ordered this month+location+acquisition channel).

WeChat data is siloed

WeChat mini programs must be thought of as a component of an omni-channel strategy, online and offline, cross-applications and cross-devices.

You need a platform to blend data from a multitude of touch points. By simply using the native WeChat platform, you won’t get anything close to that.

A traditional marketing suite will give you the tools to blend your data across channels, WeChat reports won’t.

Conversion funnels are not enough to understand navigation journeys

Part of data analysis is exploratory. Segmenting, slicing, finding correlations and making sense out of the unexpected. For this type of work, funnels can’t (and are not supposed to) compete with user flow reporting.

In WeChat native reporting, if users leave your manually configured funnels, you won’t know where they end up their journey, if they came back to the previous page and end up in a navigation loop.

In short, funnels are not user flows - and you don’t have any reports close to that.

How can I do better?

Calling data from the WeChat data API to visualize raw data on Tableau, using backend data and blending data based on unionID… there’s a number of possible solutions but none of them answer all of the points mentioned above.

After months of frustration, Wiredcraft put together a way to use your marketing suite on WeChat as if it was a website or a mobile app. On Google Analytics for example, you will be able to bridge your WeChat data with all your other properties (including goals and segment imports).

A few reasons why I think it makes sense for any product owner or data analyst:

  • real time campaign tracking based on UTM parameters
  • real geolocation based on user IPs and not WeChat account info
  • userID reports that can be used with openID, unionID or your membershipID (based on what your digital touch points are)
  • user flows that show you all possible scenarios
  • easy connection to your dataviz platforms. Use the Google OAuth to connect to Tableau, Data Studio or any other platforms

And my favorites for the end:

  • limitless segmentations (by device, dataLayer values, geolocation specifics, behavior, acquisition and so on)
  • import your segments, goals, conversions and filters segments from your analytics properties for WeChat
  • use the Google Reporting APIs for storage, ETL and DMP export.

Reach out to the marketing team to get a demo and have an in-depth view of our product.

Posted on December 11, 2018 in Data

Share this post

Scan to open in WeChat

Stay tuned

Follow our newsletter or our WeChat account; every other week you'll receive news about what we're up to, our events and our insights on digital products, omnichannel, cross-border ecommerce or whatever the team fancies.

  • Add us on WeChat

  • Subscribe to our newsletter