Read the transcription of the episode:
Data-driven decision making
In today’s business, the way decisions are reached is tremendously important. Let’s try and look into the process in our heads, teams, and the whole organization that leads to making these mature choices, often strategic for the company growth.
Recently, more and more has been said that business decisions should be supported by data. The English phrase, data-driven, suggests that they should be almost led or propelled by data. This has become more than just a slogan, especially in the Internet-based or, more widely, the Internet-related tech businesses, that often deal with massive amounts of data. The skilful use of this resource gives young startups a competitive edge over established corporations who may have accumulated more data, but aren’t always able to exploit this advantage. Listen to this episode to learn more what are the key elements worth considering while becoming a data-driven organisation.
It is often said that it is hard to correct anything without measuring, and it’s surely hard to tell if our correction worked and to what degree. We might intuitively sense positive or negative changes, but our intuition is not enough when we want to measure the improvement according to parameters and estimate its influence on the whole business.
It is also worth remembering that even measuring is not enough. Businesses just don’t get better from looking at numbers and indicators. The numbers we obtain and the analyses that we or our specialized units perform, are just an input for conclusions that must be drawn. Only after these conclusions and the resulting recommendations are implemented, improvement chances may begin to take shape. So, keep in mind, just watching numbers can’t boost your performance indicators.
By the way, we often fail to see how important it is for growing organizations to keep a good mix of competencies in teams. Isn’t it useful when someone casts away guesswork and opens Excel to run a simple analysis of something that seems so trivial but is not so easy to others? Intuition and business instinct must be complemented with solid analytic skills.
Please note, however, that the value of an analytic calculation lies its simplicity. There is a principle “Done is better than perfect”. It means that if we succeed in doing something nearly perfectly, it is better than applying lengthy examinations, that hold the project up beyond reason, trying to pursuit absolute flawlessness. The cost of medication can’t exceed the expected benefit. Pondering over a solution shouldn’t take longer than its implementation and consequences. As an example of a simple and efficient analysis, particularly in the beginning of a decision process, I recommend my podcast titled “Multitude of ideas?”. It shows how uncomplicated an analytic tool can be. Sometimes it’s just a question of cross-referencing some numbers and sorting some results according to calculated priority score. Just so little brings us so much closer to being data-driven decision makers.
Obviously, our input numbers or output decisions aren’t always that simple. The more complicated the case, the more comprehensive the analysis.
It is important, however, that despite their complexity, analyses should be made transparent and widely available to all team members. With efficient data distribution and goal delegation systems, they will find it easier to make data-driven decisions for their organization. To help them, there are BI (business intelligence) tools (such as Periscope) which visualise data much faster than it can be done manually. Not only do they visualise data – they also allow for changing basic criteria and variables, which can be defined in particular graphs, for any data sector. They allow for various segmentations and aggregations of daily, weekly and monthly data, as needed. Other tools, such as Deep BI or programming languages as Luna, facilitate the communication between the analysts and the business data users. They are very useful as they speed up the availability of particular analyses and reduce the response time. If a Board of Director member or a manager knows that a response to his inquiry to the Data Department will arrive in two weeks, he will get discouraged and ask fewer questions. With good tools at hand, such as Deep BI, Luna or Periscope, they will be able to verify lots of their hypotheses within a few minutes. Consequently, as one question leads to another, they will feel encouraged and come up with a lot more questions. Having to wait for too long, they would never do it and the data resources would remain unused.
The business intelligence tools do not only give us cool graphs and charts. What matters, is that they make the problem-spotting and answer-finding process far more informative for the team members and the whole organization. Another process that is facilitated by the business intelligence tools is the sanity check. We know that, whatever input is loaded into Excel, it will produce some result. However, only if we cross-reference this result and try to obtain the same outcome using a different set of input data, we can verify whether our approach is correct.
To illustrate this:
if we analyse our idea using raw, basic data and arrive at more general, larger-scale conclusions, this bottom-up procedure should be then verified by the top-down analysis, starting from macro indicators for a given country, market or region and moving down.
If these two – top-down and bottom-up procedures – yield comparable values, we can say we have performed a sanity check or verification. If our analyses are not subjected to counter-analyses, we might just as well not perform any analytic steps, because we may end up worse off, misguided by premature conclusions.
It’s interesting here to have a look at the evolution in the approach to data that took place at AirHelp, which I used to co-lead as its Polish CEO and as a member of its international Management Team. Its Date Department started from one, then grew to a few and finally up to 13 specialists linked to team leaders who were trained at least in a number of basic business cases. The training workshops, which I sometimes personally led, taught how to find relevant business cases to justify the viability of decisions. Those decisions, technical or non-technical, as well as process innovations, did not require any IT developments, but were vital to the company operation.
In my view, if data finding and analysing practices are successfully implemented in a company, together with the awareness of how beneficial data access and analysis can be, it will bear fruit in lots of aspects of corporate life. The transparency of data access will increase the motivation of the whole staff. They will get the sense of the company direction, its strengths and weaknesses, as well as of what things need doing so their organization can do better. This database, or data-driven awareness is indispensable not only for thee Data Department. Analysts alone cannot possibly have a hands-on experience of every company sector. They just examine the data and interpret what they see, sometimes correctly, sometimes not. Only with the competent support of the people focused on the area that is to be affected by a decision, as well as the statisticians and data engineers, can appropriate conclusions be drawn and business decisions made.
Here, it would be useful to have a look at how one of our foreign customers, a great e-grocery company, decides about entering new markets. Their process of expansion market analysis is organized in a very interesting, data-driven way. They do not arrive at their expansion recommendations just relying on their data analysis. As a matter of fact, their ultimate decisions are waiting until multi-aspect expertise on the relevant market is consulted and cross-examined with independent specialists. As you can imagine, a deployment decision is a crucial one for any company, and it seems impossible nowadays to take it without prior top-down and bottom-up data processing.
Bottom-up, starting from the data gathered in our fields of operation and then extrapolating from our observations to new markets. Top-down, using macro data, global and international analyses of export, import, consumer and other markets to cross-reference if our assumed growth or market presence is going to bring the expected results. A lot of other milestone decisions just cry for these top-down and bottom-up procedures.
give it a serious thought if your company decision-making processes are well founded on data. Could the things you do be done better or looked at from a different angle? Maybe it is not necessary to outsource chartered analysts right away. Try and find some people who have an analytical mindset, but operate in other business areas or models, where data have been used in a different way or on a different scale. They might be able to notice what your company members often fail to see.
Well, to conclude this short introduction on analysis and being data-driven, I’d like to urge you once again to derive your decisions from data, as much as you can. Nowadays, it nearly always is a key issue in tech companies, and I hope that I have made it evident to more of you. It is important not only to measure your KPIs (key performance indicators), not only keep an eye on numbers, but draw conclusions from them, that develop the company. As I told you in the beginning, the results will not get better just from staring at numbers.
If the data you have accumulated, analyses you have conducted and the conclusions you have came up with are wisely and widely distributed among your crew, they are very likely to become better motivated and help your business to flourish.
Cheers for now, good luck!.
Click the picture to listen more or record a voice message.