With discussions abounding around the power of big data and deep learning algorithms, you could be forgiven for assuming that 1) we now have the technology to measure everything and 2) everyone is doing it. But before considering whether QuintilesIMS, your CRM or Twitter can provide your company with the largest data sets, perhaps you can take a moment to ask: “What is it for?”
Historically, pharma companies have suffered from two opposing “measuring diseases”. The first afflicts the organization not tracking data aside from basic financials. In this scenario, decisions are made based on unproven ideas or gut feelings, a handful of conversations, or who complains the loudest. This method may have been sufficient to sustain a business and career once upon a time, but this is a deteriorating condition.
The other “measuring disease” is more like a paralyzing anxiety, easily recognized by the organization drowning in numbers that few people look at with discernment. This usually happens when you have a good intention, a fistful of data sources and overly creative ways to calculate numbers. The intention is a good one: make decisions based on data that reflect the status of the business and trends and signals that could indicate shifts to react to. If the data is good, it’s great. But if it’s bad, it can lead somewhere between nowhere and a disaster. Creativity has a place, but not in churning out calculations of the same thing. What could be missing is a bit of understanding and some conscientious editing.
Before letting a business intelligence (BI) or reporting project take hold of the organization with a ton of metrics (often painstakingly organized into reports and possibly customized to various users), take time to invest in the right design:
- First, make sure that you understand the strategy of the company and the tactics to be implemented that support the strategic objectives.
- Second, you need to know what actions you are placing in the market (e.g., a new customer list for reps, promotional messages, concerted promotion in different channels) and what you’re expecting to get out of those actions (e.g., an increase in call activity with certain segments, attendance at promotional events, change in brand perception). Hopefully any proposed activity will translate into an increase sales. This is an important concept to keep clear: measuring sales tells you what has happened in the past, but you track what you’re doing to understand whether it is working to get you more sales in the future. This understanding of what is working is what most business intelligence programs miss.
How Key Performance Indicators (KPIs) can help business intelligence
Figure 1 below illustrates some of the points that we believe are essential in planning for your Key Performance Indicator (KPI) definitions. Even if you have already defined a ton of metrics, try to go through a similar process and then bring in your metrics to see whether they fit. Ultimately, what distinguishes a KPI from a metric, is its ability to measure what you’re trying to execute or the result you’re trying to achieve, in a way that can preferably answer a question and at least generate a hypothesis, leading to an actionable decision. So, before a metric makes it as a KPI, make sure it passes the following assessment:
- Does it measure something you need? (It must be relevant to the decisions you’re trying to make.)
- Does it inform? (It must support or contradict your hypothesis.)
- How accurate is it? (It’s ok if it needs other metrics/KPIs to complement it, but it should add to your understanding. If it’s neutral in value of information, get rid of it!)
You’ll notice that we don’t actually define specific metrics or KPIs. The reason is that the focus should be on what needs to be understood, followed by the best way to measure it. A lot of business intelligence programs fail because they start with what they have, in terms of data or tools, and define different ways of displaying versions of that data, rather than starting with what’s actually needed. This often leads to overly-engineered reports and the production of metrics that are not genuinely informative or insightful.
Certainly, the source, quality and types of data are absolutely relevant to defining KPIs, but the first step is still crucial to define these three aspects of data. The handling of data is itself a complex aspect, and will be the topic of our next blog post. You may also notice that there isn’t a linear relationship between one measurement and one conclusion. This is key to point out – true understanding includes complementary data points that support a view of what is happening. A single data point rarely contains sufficient information.
Additional points of discussion
Some additional process pitfalls can occur in the handling and reporting of data:
- Cost – This element can be increasingly difficult to justify, but only if the value of the new intelligence on future decisions isn’t demonstrated.
- Accuracy – How closely certain data types can reflect with accuracy what is happening in the market, is another concern. The best way to handle this issue is by accepting the limitations and reflecting them in the answers your BI produces. If data is only accurate by proxy, make sure to find other, supporting data points for any conclusion and/or the data evidence that would refute it. Finally, if there is still uncertainty, use scenarios that show the consequences of different decisions or paths taken when the uncertain measurement is taken to extremes.
- What and how you measure – Another point to question is what and how you measure things, locally vs. centrally. The one statement we will make here about that discussion is that it is far more valuable to have simple metrics that can be understood and to take action on, than to have complex ones that few understand or find useful. This illustrates another important facet of KPIs – they not only inform what is happening but are also a useful tool to direct an organization. Visionary leaders must identify the KPIs that most closely direct their teams to follow their vision and will often socialize these single KPIs in their communications and incentives. This is less about having a complete vision of what is happening in the market and more about focusing sales teams on the right objectives.
Business intelligence is an essential element of achieving commercial excellence. It is a process of educating teams rather than a technology implementation project. Spend more time understanding what the business needs. The data and technology to handle it will follow.