Marketers are under extreme pressure to be data-driven. Yet, you won’t find marketers talking about poor data quality or questioning the lack of data management and data ownership within their organizations. Instead, they strive to be data-driven with bad data. Tragic irony!
For most marketers, problems like incomplete data, typos, and duplicates are not even recognized as a problem. They’d spend hours fixing mistakes on Excel, or they’d be researching for plugins to connect data sources and improve workflows, but they are not aware that these are data quality issues that have a ripple effect across the organization resulting in millions of lost money.
Marketers today are so overwhelmed with metrics, trends, reports, and analytics that they just don’t have time to be meticulous with data quality challenges. But that’s the problem. If marketers don’t have accurate data to begin with, how in the world would they be able to create effective campaigns?
Here are his insightful answers to my questions.
- What were your initial struggles with data quality when you were building your product? I was setting up a review generation engine and needed a few hooks to leverage to send review requests to happy customers at a time when they would likely leave a positive review.
To make this happen, the team created a Net Promoter Score (NPS) survey that would be sent out 30 days after signup. Whenever a customer would leave a positive NPS, initially 9 and 10, later expanded to 8, 9, and 10, they would be invited to leave a review and get a $10 gift card in return. The biggest challenge here was that the NPS segment was set up on the marketing automation platform, while the data was sitting in the NPS tool. Disconnected data sources and inconsistent data across tools became a bottleneck that required the use of additional tools and workflows.
As the team went on to integrate different logic flows and integration points, they had to deal with maintaining consistency with legacy data. Product evolves, which means product data is constantly changing, requiring companies to keep a consistent reporting data schema over time.
- What steps did you take to resolve the problem? It took a lot of working with the data team to build up proper data engineering around the integrations aspect. Might sound pretty basic, but with many different integrations, and lots of updates shipping, including updates affecting the signup flow, we had to build a whole lot of different logic flows based on events, static data, etc.
- Did your marketing department have a say in resolving these challenges? It’s a tricky thing. When you go to the data team with a very specific problem, you might think it’s an easy fix and it only takes 1h to fix but it really often involves a ton of changes you’re not aware of. In my specific case regarding plugins, the main source of problems was maintaining consistent data with legacy data. Products evolve, and it’s really hard to keep a consistent reporting data schema over time.
So yeah, definitely a say in terms of the needs, but when it comes to how to implement the updates etc. you really can’t challenge a proper data engineering team who knows they have to deal with lots of changes to make it happen, and to “protect” the data against future updates.
- Why aren’t marketers talking about data management or data quality even though they are trying to be data-driven? I think it’s really a case of not realizing the problem. Most marketers I’ve talked to widely underestimate the data collection challenges, and basically, look at KPIs that have been around for years without ever questioning them. But what you call a signup, a lead, or even a unique visitor changes massively depending on your tracking setup, and on your product.
Very basic example: you did not have any email validation and your product team adds it. What’s a signup then? Before or after validation? I won’t even begin to go into all the web tracking subtleties.
I think it also has a lot to do with attribution and the way marketing teams are built. Most marketers are responsible for a channel or a subset of channels, and when you sum what each member of a team attributes to their channel, you’re usually around 150% or 200% of attribution. Sounds unreasonable when you put it like that, which is why nobody does. The other aspect is probably that data collection often comes down to very technical issues, and most marketers aren’t really familiar with them. Ultimately, you can’t spend your time on fixing data and looking for pixel-perfect information because you just won’t get it.
- What practical/immediate steps do you think marketers can take to fix the quality of their customer data?Put yourself in a user’s shoes, and test every single one of your funnels. Ask yourself what sort of event or conversion action you’re triggering at each step. You’ll likely be very surprised at what really happens. Understanding what a number means in real life, for a customer, lead or visitor, is absolutely fundamental to understanding your data.
Marketing Has the Deepest Understanding of the Customer Yet Struggle to Get their Data Quality Problems in Order
Marketing is at the heart of any organization. It’s the department that spreads the word about the product. It’s the department that is a bridge between the customer and the business. The department that quite honestly, runs the show.
Yet, they are also struggling the most with access to quality data. Worse, as Axel mentioned, they probably don’t even realize what poor data means and what they are up against! Here are some stats obtained from the DOMO report, Marketing’s New M.O., to put things into perspective:
- 46% of marketers say the sheer number of data channels and sources has made it more difficult to plan for the long term.
- 30% senior marketers believe the CTO and IT department should shoulder the responsibility of owning data. Companies are still figuring out ownership of data!
- 17.5% believe there is a lack of systems that collate data and offer transparency across the team.
These numbers indicate that it’s time for marketing to own data and demand generation for it to be truly data-driven.
What Can Marketers Do to Understand, Identify, and Handle Data Quality Challenges?
Despite data being the backbone for business decision-making, many companies are still struggling with improving their data management framework to address quality issues.
In a report by Marketing Evolution, more than a quarter of the 82% companies in the survey were hurt by substandard data. Marketers can no longer afford to sweep data quality considerations under the rug nor can they afford to be unaware of these challenges. So what can marketers really do to address these challenges? Here are five best practices to get started with.
Best Practice 1: Begin to learn about data quality issues
A marketer needs to be as aware of data quality issues as their IT colleague. You need to know common problems attributed to data sets which include but are not limited to:
- Typos, spelling errors, naming errors, data recording errors
- Issues with naming conventions and the lack of standards such as phone numbers without country codes or using different date formats
- Incomplete details like missing email addresses, last names, or critical information required for effective campaigns
- Inaccurate information like incorrect names, incorrect numbers, emails etc
- Disparate data sources where you’re recording information of the same individual, but they are stored in different platforms or tools preventing you from getting a consolidated view
- Duplicate data where that information is accidentally repeated in the same data source or in another data source
Here’s how poor data looks in a data source:
Familiarizing yourself with terms like data quality, data management, and data governance can help you go a long way in identifying errors within your Customer Relationship Management (CRM) platform, and by that stretch, allowing you to take action as needed.
Best Practice 2: Always Prioritize Quality Data
I’ve been there, done that. It’s tempting to ignore bad data because if you were to really dig deep, only 20% of your data would be actually usable. More than 80% of data is wasted. Prioritize quality over quantity always! You can do that by optimizing your data collection methods. For instance, if you’re recording data from a web form, ensure you collect only data that is necessary and limit the need for the user to manually type in the information. The more a person has to ‘type’ in info, the higher they are likely to send in incomplete or inaccurate data.
Best Practice 3: Leverage the Right Data Quality Technology
You don’t have to spend a million dollars on fixing your data quality. There are dozens of tools and platforms out there that can help you get your data in order without kicking up a fuss. Things these tools can help you with include:
- Data profiling: Helps you identify different errors within your data set such as missing fields, duplicate entries, spelling errors etc.
- Data cleansing: Helps you clean your data by enabling quicker transformation from poor to optimized data.
- Data matching: Helps you match data sets in different data sources and link/merge the data from these sources together. For example, you can use data match to connect both online and offline data sources.
Data quality technology will allow you to focus on what matters by taking care of the redundant work. You won’t have to worry about wasting time fixing your data on Excel or within the CRM before starting a campaign. With the integration of a data quality tool, you’ll be able to access quality data before every campaign.
Best Practice 4: Involve Senior Management
Decision makers in your organization may not be aware of the problem, or even if they are, they are still assuming it’s an IT problem and not a marketing concern. This is where you need to step in to propose a solution. Bad data in the CRM? Bad data from surveys? Bad customer data? All of these are marketing concerns and have nothing to do with IT teams! But unless a marketer steps up to suggest solving the problem, organizations may do nothing about data quality issues.
Best Practice 5: Identify problems at the source level
Sometimes, poor data issues are caused by an inefficient process. While you can clean up data on the surface, unless you don’t identify the root cause of the problem, you’ll be hit with the same quality issues on repeat.
For example, if you’re collecting lead data from a landing page, and you notice 80% of the data has an issue with phone number entries, you can implement data entry controls (like placing a mandatory city code field) to ensure you’re getting accurate data.
The root cause of most data problems is relatively simple to solve. You just need to take out time to dig deeper and identify the core issue and make the extra effort to solve the problem!
Data Is The Backbone Of Marketing Operations
Data is the backbone of marketing operations, but if this data is not accurate, complete, or reliable, you’ll be losing money to costly mistakes. Data quality isn’t limited to the IT department anymore. Marketers are the owners of customer data and therefore must be able to implement the right processes and technology in achieving their data-driven goals.