Leveraging corporate data assets is key to achieve strategic goals

Our analytical methodology, out-of-the-box solutions for managerial decision-making

Our company has developed an analytical methodology that allows us to present the company and its activities from a previously unseen perspective by examining the available corporate data assets, thus helping management to achieve its strategic goals. We find patterns and correlations that are not otherwise obvious or demonstrable, and summarise these to formulate situation assessments and recommendations that are supported by evidence and often lead to out-of-the-box solutions.

The subject matter of analysis can be very diverse; we can analyse a company's product portfolio, its customers, its pricing, the efficiency of its value-creating processes, its costs, the performance of its subcontractors, the effectiveness of its purchasing, sourcing, anything that is at the forefront of its corporate strategy - with little exaggeration, if there is data, we can promise an effective solution or situation analysis for any problem.

How do we design the data analysis?

1. The key to success lies in the research design

In the first stage of the work, we need to understand the management expectations and strategic questions we need to answer, and in the light of this, we start to draw up a research plan, setting out what data we ideally need to achieve the best results, and what research methods and analysis we will use to reach our conclusions. The plan is at this point just an itinerary. What we will actually do is basically determined by the availability and quality of the available data; if there is insufficient data, we will try to improve and enrich the data set by modifying the test methods, simulations, similar data sets, etc., but often we will also consult with management to achieve better results with minor changes to the question set.

Once it is clear what data we want to analyse, we finalise the research design, which is from then on the tool for documenting our work; in all cases, we work in such a way that our clients can reproduce our analyses on the basis of the documentation provided, even extending them to other areas.

2. Data cleansing and data quality: the foundation on which to build

As soon as we receive the requested data - that is, as soon as we start receiving it, since we often need non-trivially retrievable data that is not always readily available - our first task is to check its quality and, if necessary, bring it to a state suitable for analysis; we remove empty or useless data, standardise formats, cell types, filter out duplicates, check and validate strange or outlier values. Data cleaning work is validated with the client's experts and documented in detail for reproducibility.

3. Engaging external data sources: how do they enrich our analysis?

We almost always enrich our corporate data assets with external data sources; what we use depends on the brief, we often use KSH and OPTEN data, but any available source of proprietary data can be used, price lists, chamber of commerce lists, demographic data, geographic data, etc., depending on the research design.

4. Profiling: the role of historical data and behavioural analysis in strategy making

It is common in analytical work to profile a product, a customer group, a behaviour, in order to present the company's position through it, thus helping strategy-making: in this case, we infer the characteristics of a set of attributes based on past data and behaviour. Ideally, we have enough descriptive data to build so-called hard profiles, which are defined on the basis of quantitative data.

However, it is often the case that we do not have enough data points, in which case we usually use so-called soft profiling; in this case, we use a focus group expert workshop to identify the properties and descriptors that are characteristic of the entity we are looking for, and thus search for descriptive data for them. Soft profiling is often useful even when there is sufficient data for quantitative analysis, as it is a great validation tool and helps to make our analysis more accurate.

Stories behind the graphs: analysing data and their strategic message

Our analyses are more than a mass of graphs. Data and its representation - in graph or infographic form - tell, illustrate or prove a story. Because we are looking for answers to strategic questions, our message in our analysis needs to be clear on every page, on every slide, followed by analysis and conclusion.

A constant demand from management to analysts is that I understand the data, but what does that mean?

We believe that analysis is more than just a list of values: our job is to show correlations, patterns, examine and explain correlations, cause and effect, understand and explain the reasons for outlier data points; often an unexpected correlation or examination of data will highlight important trends.

If we do not have a clear explanation or conclusion, we also indicate that our studies may have much to offer, but they may not be sufficient, the data to explain them may not be available.

In examining data and correlations, we draw heavily on the industry experience of our experts, and in addition to the methodology, we can bring in considerable sector-specific knowledge in the fields of infocommunications, banking, healthcare and pharmaceuticals, as well as in certain areas of precision manufacturing and agriculture.

Combining AI-based models with expert analysis for deeper understanding

Business analytics is a very powerful tool in the hands of an information-hungry business leader when strategizing, but it is not our most powerful tool. A business analyst is just one person, with finite computational and analytical capacity but considerable interpretive ability.

Recognizing this - and true to our company slogan: "One source, One Partner for Data & Consultancy" - we can combine our traditional analytics with machine learning and AI-based models on demand.

In this way, we can take all measurable factors into account together, but also use expert analysis to make it easier to grasp the relevant indicators, validate them and provide clients with a logical expert interpretation of the model results.