Drowning in Data

(Originally published in German Chamber of Commerce Greater China’s Ticker magazine Spring 2021 edition| Data and Marketing Trends in Fast-Paced China)

You can’t manage what you can’t measure. The old business adage by Peter Drucker is overused, but it holds up. With each new campaign launched, social marketers collect an astounding amount of information on how their audience responds to their content. This has raised increasing concerns for privacy intrusion, amplified by the emergence of data privacy laws such as the GDPR or the Cambridge Analytica scandal. Yet, while marketers are drowning in data, we see in reality that teams are just not data savvy.

Data is the single most powerful tool your marketing team has to take on the China social media landscape — that is, if they are equipped with the tools and techniques to use this data to their advantage. The competition to capture your target audience’s attention is getting fierce — marketers are faced with dwindling resources and dropping open rates.

Yet, social media is an inexpensive channel for any brand to quickly iterate and refine their brand positioning. Marketers need to stop shooting content from the hip and start to think of themselves as snipers. This is where data-driven decisions are needed the most.

How Can Data Boost Social Marketing Efforts?

Great social marketing starts with data. There is no universal shortcut, industry trend of ‘best practice’ that will lead to success.

Data must drive your marketing teams’ efforts. The most basic analytics can reveal a lot on what type of content your audience enjoys, how it compares to rivaling brands, or when is the best time to post.

While marketing teams are drowning in data, very little of it seems to be used for insights. A clear example is the pattern of content published over Chinese New Year. China marketers tend to post a large volume of their WeChat content right before the Chinese New Year holiday, yet the data shows a significant drop in read performance then. If marketing teams would take note of these trends revealed by data, they’d know that content performs much better in the second half of the Spring Festival weekend (see fig. 1)

Fig. 1: Data based on analyzing 12,770 WeChat articles from 264 Official Accounts, published during the year of 2020. Design: Alex Duncan

Rather than subscribing to common industry practices, marketing teams must invest time and resources into analyzing their own data, which may hold the key in significantly improving marketing efforts.

What Data Do I Need to be “Data-Driven” ?

More data is not necessarily better — or even possible. In reaction to recent concerns about user privacy and data collection, platforms are increasingly restricting access to data — in China even more than in the West.

As personal user data is becoming harder to obtain, marketers need to turn their full attention to readily available quantitative data — how their audience interacts with their content. Even without knowing your readers’ exact location, you can still learn a lot from how your followers react to different content themes.

Even so, today’s marketing teams are swimming in data, and few will dispute that it holds incredible business potential. And yet, with 60% to 73% of all enterprise data never being analyzed (Forrester, 2016) do marketers really know what to do with it? In many cases, data appreciation does not translate into teams using it for informed decision-making. Collected data is only as useful as the humans who interpret it.

Developing Data Literacy

One of the major roadblocks to getting started is the misconception that working with data is a specialized technical skill.

A Shanghai-based Digital Transformation Manager we spoke to described how the marketing team in her organization would collect massive amounts of data each day, which they would later do nothing about. Upon closer observation, she realized the team was not equipped to clean, organize and contextualize any of it.

This is reflected in a study by Accenture, stating that only 25% of employees feel fully prepared to use data effectively. However, most companies do not necessarily need to hire a data scientist — their existing team usually possesses the essential skills. It might be hard to believe, but most of what we learned in math class at high school is suited to analyze data.

Companies also don’t need complicated software. A simple spreadsheet is more than enough for the majority of marketing data analysis. Modern spreadsheets can handle thousands of rows, each containing hundreds of data points. They also have a wealth of formulae, and a wide range of ways to visualize the stats.

It’s often helpful to start with a simple frequency distribution to get an overall picture of the data. Be clear about when to use a median average instead of the mean (see Fig. 2). This will help avoid your averages from being skewed by posts that were promoted or went viral.

Don’t expect this process to be quick and easy. Allocate time and space to learn and experiment. You may also consider hiring a consultant or a data specialist to help get your team started.

Visualizing data doesn’t necessarily have to be about graphs. Modern spreadsheets have simple ways to turn a dull table of numbers into a colorful map where you can easily spot patterns. Once you’re starting to generate results, be sure to question the data.

No data set is ever clean, so if you do spot anomalies, go back and look at the underlying data you collected. It can be as simple as a human error when collating the raw data, or an outlying, once in a blue moon event, that doesn’t represent the overall trend but significantly disrupts the picture. Be careful of drawing conclusions from too little data. One post about a topic isn’t necessarily enough. Test ideas or themes over time to see if there is a wider pattern.

Although you can start simple, sophisticated tech might be needed at times. When data becomes too much for a spreadsheet, an SQL database might be required. At that point, you might want to consider bringing in more outside expertise or hiring a full time data analyst.

Build a Data-Driven Culture

As marketers learn the true value of data and are empowered with data-processing skills, brands will start reaping the benefits of a data-driven social marketing strategy.

To truly become data driven, companies need to build this into their decision-making processes. Make it a habit to ask your colleagues: What does the data tell us?

Challenge the team to provide data to both support and contradict their theory. Ensure that whenever data is collected and presented, it will be turned into an opportunity to draw conclusions from. Celebrate each time you find a new data-based insight to motivate your team.

In addition to transforming your existing marketing activities, data is opening new opportunities to shape the customer experience. Platforms like WeChat offer powerful ways to connect with your audience, but your team must be on top of the data before you can start.

We can be certain that new challenges will appear for marketers worldwide. Building a data driven culture will help you weather the storm, setting you in great shape for the future.

BONUS:

7 Do's and Don'ts of Data Driven Marketing to Get You Started

1.DON’T dive into data head-first without a clear objective. Do you have a clear idea on what do you want to achieve?

2. DO start by forming clear questions or hypotheses that you hope to test with data. You might have a hunch of what kind of work produces the best results — now prove it with data.

3. DON’T be afraid of being wrong. Challenging existing assumptions often results in the most impactful insights. E.g. you have always been publishing your content on Thursday evenings, when actually content posted on Monday mornings performs much better.

4. DON’T get distracted by vanity metrics. Just because a number is large, doesn’t mean it’s generating relevant outcomes to your business. E.g. WeChat article likes > no tangible outcome.

5. DO focus on the metrics that support your business objectives. For example, if your goal is to drive sales, spend time to work out what drives traffic to your e-commerce site. Or, perhaps you are trying to build a stronger connection with your customers? In this case you should focus on engagement in the forms of comments, likes and re-shares.

6. DON’T forget about the wider context. Just because a metric has changed in a certain direction, is this pattern specific to your brand or part of a wider trend across the entire ecosystem? For example, on average, Read Rates on WeChat have fallen by 65% over the past 5 years. If you’re seeing a similar decline in your read numbers, you can conclude it’s just part of a broader trend across WeChat, rather than a specific reflection on your account.

7. Finally… DO think about your audience. Although we’re talking about data, marketing is still fundamentally about people. If you have customer archetypes, this can be a great opportunity to validate them, e.g. your customer persona of a millennial Chinese girl from a tier 7 city is price sensitive. Does content focusing on price perform better?

What are some of your best tips for working with data?

(Originally published in German Chamber of Commerce Greater China’s Ticker magazine Spring 2021 edition| Data and Marketing Trends in Fast-Paced China)

References:

Gualtieri, M. Hadoop Is Data’s Darling For A Reason (2016). 

(Qlik, Accenture) The Human Impact of Data Literacy (2020).