Data-led value creation in three steps
Data has become a business division in many organizations. It validates opinions and can mean the difference between an idea being signed off or rejected. Access to reliable data can also mean the difference between success and failure.
Data sharing between marketing and product teams is a critical for growing a business. It ensures that the insights gleaned from users are acted upon by stakeholders and continual improvement is delivered to the customer experience and product interactions.
To maximize the flow of insights between marketing and product, find out which metrics are most important to each team, why, and how they are measured. It’s eye opening how something so standard as “profit” or “revenue” can take on so many definitions across job functions, departments, and companies.
Aside from different definitions for metrics, it’s also common for teams to use different tools to measure and report on performance. This often creates complexity that leads to confusion, breeding distrust in the data itself.
In one organization I found many metrics added each month to their reporting. The result? An analytics platform containing a complex and inconsistent web of metrics. Rather than enabling fast and trusted business decisions, the incongruity bred a behavior of trying to find a metric to match the desired story. Constantly changing measurement standards created mistrust in others’ analyses. Once trust in a data source is eroded, it is hard to rebuild.
So how can companies ensure their entire organization benefits from data insights through trusted, relevant and centralized data gathering?
(1) Define sources of truth
Ultimately, teams from across the business need to use the same data sources and metrics.
The first step is to define primary sources of truth:
For on-site behavior and traffic data, use a third party platform such as Google Analytics. Having a reputable third party measurement source is particularly important for publishers who must bill others on the traffic they serve.
For cost data, the source of truth should be the marketing partner who bills you. Their figures should be sanity-checked with your own analytics solution.
For revenue and conversion data, utilize your internal central database. Ultimately, no third party will be able to verify incoming revenue and orders like your own accounting team.
It is critical that consistent definitions and data collection methodologies are agreed. Not everyone across the organization will know what a conversion pixel is or how asynchronous code affects them, and maybe they don’t need to, but everyone should have the confidence that the methodology is unwavering in its consistency. Share these definitions with everyone in the company so the language is universally understood.
(2) Define backup sources of truth
We live in a world where primary tracking systems go down and have bugs more often than we would like. Knowing which system to use to verify data movements can help alleviate concerns that the house may be on fire.
If, for example, a company’s primary source of truth on traffic reports a 50 percent drop, but a secondary source is in line with expectations, that likely points to a tracking problem and not a true drop. In addition, if a backup source consistently undercounts a particular metric by 10 percent for example, when the primary source goes down, teams can fill in the gap reasonably accurately.
One last benefit of having backup sources is quality assurance. When primary and secondary sources of truth do not move harmoniously, it usually points to an implementation issue, often the case after new product releases.
It’s important to educate teams on how things are tracked and when to sound an alarm.
Employees need to develop an understanding of what performance variations are deemed acceptable. Through group sessions, effort should be made to educate employees on basic data analysis such as what standard deviation is tolerable to the organization. Anything within the range may simply be statistical noise; anything greater usually signals a departure from normal. This understanding helps focus teams on when to act on data insight or dive deep on an anomaly.
Following these steps will help put the foundations in place for a robust data measurement model. Without clear definitions and formulas, the data is all but useless. Consistent sources of truth, a reliable back up system and a team of people who are aligned all ensure that an organization can reap the value that resides in today’s world of big data.
This post was first published on Naspers