What are the most effective ways to collect and validate data in consulting?

Still love the old fashioned job shadowing end users as they do the work to not only understand the problem better but to uncover additional issues within the process.

Here's a hard truth - If you can't explain a client's problem to a child, you don't understand the problem well enough. Making sure you understand a client's goals and expectations is important. This is especially true at the outset of a consulting project before you dedicate time, resources, and money to solving a problem. Otherwise, you might find that you've done an entire consulting project without actually solving a client's problem.

In my experience, most clients have great difficulty defining the problem correctly. Consultants are in a unique position to facilitate this discovery process, bringing the conversation closer to the client's perspective and direct experience. Documenting the current state and the desired future state is a meaningful, informative, and collaborative first step that it's not only an essential foundation for problem definition, but also achieves broad consensus and suggests a journey.

I am always listening for clues, that a client might have a problem that they are unwilling to admit, presents ego challenges, or is beyond the scope of that they have considered. This is a sensitive process, but many clients have psychological barriers that prevent them from seeing what is really going on, accepting risks they should take, or having others give them advice. I have found academic tools and practices less effective, than building sufficient report with clients that give them trust in your guidance. One way of achieving this is to share your own experiences and business challenges, and how they made you feel, and the new decisions you had to make. This sets the stage for more candid conversation about the matters at hand.

Once you define the problem, come up with the best hypothesis of the answer to the problem - e.g., "We can increase profits by focusing on increasing revenues instead of cutting costs." Then you break the hypothesis down into 2-5 assertions that need to be true to support that hypothesis - e.g., "Revenues can be increased." and "Costs cannot be cut.") Test those assertions to make sure they completely and uniquely support the hypothesis. Then break those big assertions into sub assertions - e.g., Revenue could break down into "Increase Volume" and "Increase Price".) You break the problem down until you can think of the slides that would prove/disprove each assertion. That tells you exactly what data you need and avoids wasted effort.

There are four kinds of data sources to consider. - Quantitative - Qualitative - Primary - Secondary - Quantitative data sources include numbers. For example, quantitative data can include budget line items, growth rates, forecasts, headcount, number of competitors, inventory levels, financial ratios, and more. Qualitative data sources may not involve any figures or numbers. In consulting projects, these data often come from interviews. Primary data sources are close to the source. It's not hearsay but direct information. Secondary data sources are removed from the project or your own data collecting. The best data is from unbiased, current, accurate sources, regardless of the source or type of data.

The next step is to identify the data needed for validation, understanding, analysis, testing, and measurement. This includes not only structured data, but also sources where almost all of the data is unstructured, such as interviews, open-text surveys, and observations. In this Data Engineer role, the consultant evaluates data samples for completeness, accuracy, and relevance. No data source is perfect, but data wrangling should be kept at a minimum. In my experience, bad data is rarely successfully cleaned up during a consulting engagement, leading to awkward situations like pitching prototypes to key stakeholders using inaccurate data. This type of situations generates mistrust. Low-quality data, beyond salvation, should not be used.

The best way to ensure you are collecting the right data, and only the right data, is to create the "blank slides" that the data will go into. Each blank slide will have the statement it is trying to support (or disprove) at the top and a chart of something without the data in it. Collectively the blank slides will be all the slides in the final presentation. It is a great way to focus data gathering and analysis on just what is needed and avoiding getting into the "boil the ocean" mode down data analysis rabbit holes chasing data that is interesting but tangential and not essential.

Here are some tips for selecting the best data sources: Define your research question Identify relevant data sources Evaluate data quality Consider data freshness Assess data reliability Check for bias Evaluate cost and accessibility

Here are 7 reasons why this is a difficult step in the process: - If the data was easy to get, the problem would be solved - Lots of data is not structured (e.g., customer perceptions) - There is a lot of "tribal knowledge" in people's heads - The data is often in multiple IT systems - The data is rarely clean; too often, "junk in = junk out" - Your data request is not the IT analyst's priority (trust me on this) - Too often, data tells about lagging indicators about the past

Being a consultant looking for data can be a bit like a scavenger hunt. You know that someone somewhere at the client organization has the answers to the questions you need. The catch? You have to find them! Good record-keeping is part of good data hygiene, and it can be really helpful when you're on a consulting scavenger hunt for data and insights.

Ironically the most important skills in gathering data are often interpersonal, not analytical skills. You often have to convince busy folks at your client to do the extra work beyond their normal job to gather data for you. Learning to be clear, concise, persuasive (and politely pushy) are key to being a great data gatherer.

You will never be able to collect ALL of the data. Avoid "boiling the ocean". When it comes to data, it's important to focus on the most relevant and meaningful information. Collecting and analyzing data can be a valuable tool for decision-making and problem-solving, but only if the data is relevant and properly analyzed. To avoid boiling the ocean with data, it's important to start with a clear goal or question in mind, and then collect and analyze data that is directly relevant to that goal or question.

The key phrase here is - Measure twice, Cut once. Before laying everything on the line in a recommendation to your client (or even when performing analysis), you need to confirm that the data are good and clear. As in all things consulting, assurance, and audit - "Trust but verify." For quantitative data, a second or third set of eyes is good. For qualitative data, it helps to confirm what you believe are critical findings with multiple parties most likely to be "in the know."

Assume your data is wrong until you prove it otherwise. There are going to be people at the client who will attack your analysis because it makes them look bad. Imagine a scenario when a worried client tries to create confusion or delay by challenging your data and analysis. You will want to be able to say with COMPLETE confidence that the data and analysis are right. Check the spreadsheet math by doing it roughly in your head. Randomly and manually check cell references in spreadsheets. Find other numbers that should have a relation to the number you calculated to see if it jibes. Make it a habit of doing the "what is the average per unit" math in your head when you hear a total. Get someone else to look at it. Do whatever it takes.

Validating data is crucial for ensuring that it's reliable and accurate. To validate data, check it for accuracy and completeness, cross-check it against multiple sources, validate it against benchmarks or standards, use statistical methods, review and correct errors, test it in real-world scenarios, and seek feedback from experts.

The most important things to keep in mind at this stage are the following two: 1. Why are we doing this? One should know the intent and objective behind the analysis, and accordingly prioritize insights generation 2. Are there any outliers skewing my findings or outcomes? This is supercritical. More often than not people end up blindly using averages that include the outlier data points as well. For a reliable actionable insight, it is very critical for one to have a closer look at data, and identify outlier cases that can be ignored out-right (and probably be mentioned in the footnotes).

Collecting data is something you do with a goal. The purpose of collecting, validating, and cleaning data is to tease out valuable implications that can be the bedrock of meaningful analysis driving to insights that are supportive of subsequent action.

Avoid "analysis paralysis" by resisting the temptation to do analysis that is interesting but not essential to solving the specific problem you are working on. If that "interesting but not essential" line of analysis turns out to be essential after all on something later, it will still be there then.

Make sure to avoid some common mistakes with analyzing data: Not defining clear objectives: Without clear objectives, it's easy to get lost in the data and end up drawing incorrect conclusions or insights. Using biased data: Using data that is biased or incomplete can lead to flawed analysis and decision-making. Ignoring outliers: Ignoring outliers can lead to incorrect conclusions and insights.

It's critical to be a good storyteller here and to include the impact of data that you've uncovered in a manner that defines the business outcomes.

Having good data isn't enough. You need to tell a compelling story with your data. This is critical if you want to get buy-in for your analysis and any recommendations that flow from your analysis. You most certainly want that! After all, it's the goal of a consulting project to provide a compelling analysis that leads to meaningful action.

Less is more when it comes to presentations. Start with a one page Executive Summary that clearly defines the problem, the recommended solution, and the key analysis/data that show that the recommendation has to be the correct one. Having the whole story up front helps your audience identify the key questions in their head that need to be answered. After the exec summary slide, you should only have slides that tie back to bullet points on your executive summary. Keep it short. Avoid interesting tangents that may distract. Every slide after #10 is probably taking focus off your key conclusion. If you have done your job well, you have the analysis and answer to those questions in your short slide deck.

The data you analyzed should tell a story to the client no matter how technical/analytical they are. Make sure you simplify the data, highlight key findings, provide context, know your audience, and anticipate questions!

'Not everything that can be counted counts and not everything that counts can be counted' (attributed to Albert Einstein)and Data collection and validation are crucial steps in the problem solving cycle in consulting however there is data and data. Data and statistics are key, BI too, however, we need all that intangible uncomfortable truths that often cannot be easily captured, but make all the tangible impact and make the difference between success and failure. Spend time really being there, looking at the client, the said and the unsaid, because 80% of the culture lives in the assumptions, inside the hearts and heads of the leaders, and the workforce, and you will be a great management consultant!

The most popular analytical tool for management consultants varies depending on the specific industry, problem, and data being analyzed. However, some of the most commonly used analytical tools for management consultants include: Excel: Excel is a widely used tool for data analysis, modeling, and visualization. SQL: SQL is used for querying and manipulating large datasets. Tableau/PowerBI: For data visualization and creating dashboards