Spreadsheets are the most used, and most misused tool when it comes to business analysis. Through the built-in capabilities, creative use of functions, and grit of users, Excel has provided the backbone of strategic analysis for decades. But just look at the (lack of) progress we’ve made in the past 35 years:
Given the growth in size of data and the increased depth demanded from data analysis, it might seem that analytics needs have now exceeded what Excel is able to provide. What is at the root of its staying power?
Three Reasons for Excel’s Success
Code Optional Beats No Code and Code: Excel allows many outputs and transformations to be made without the use of code, but also supports extensive formula-based functionality.
Starting from grade school, everyone is able to take Excel results and create charts, tables, and other outputs. Certainly no SQL and no Python is needed; this low onboarding cost is the first selling point of Excel. But in truth, Excel formulas are just as much code as simple transformations that can take place in Python. By allowing users to employ code directly, Excel is able to both be feature rich, and simultaneously avoid the “fighter cockpit” problem of having every possible feature become a button on a crowded interface. Meanwhile, in the data science world, no single code optional tool exists. Notebook software gives a great data science -coding experience; a completely different set of tools lets users build models with no code. However, things work better when combined together; no code is faster sometimes, while code gives infinite extensibility other times. Plus Excel prevents siloing users by level of technical sophistication; code-optional workbooks can be shared to everyone and used by anyone.
No Switching Costs: Excel is nimble, and lets you seamlessly move between transformation/wranging, visualization, and analytic results.
With integrated data summary tools, charting tools, and analytic functions, Excel individually allows a non-technical user to accomplish almost all practical tasks. A full stack of tools now exist to make every step of data analytics easy, but they all exist as segregated tools. Data wrangling (Trifacta), visualization (Tableau), bespoke exploration (Python), model building (Azure ML) are all easy. But once you slap them all together, the process is no longer as smooth, with each tool treating only one or two of the concepts as first class. While Excel is not as good as any of these tools in their best dimension, it can take all modern analytics tools on with its well-roundedness.
Seeing Is Believing: Excel is intuitive because people are always able to see their data and how it moves in each step or transformation.
If a cast does not work in Excel, the user can immediately see it and take action to resolve it. An analyst can be confident that progress is being made correctly, vet each step visually, and then think about what the next step should be while looking at the outputs of the previous step. By contrast, too many “no code data science” tools or workflow engines actually lock data behind analytic building blocks. The result is a clean analysis pipeline, but completely sanitized of the underlying data. As a result, users report that problems arise when an operation / transformation fails to do the right thing, outliers aren’t caught, or a bad filter completely ruins the usability of the workflow.
This is good, but you can do better
Our perception of Excel as the de facto business analytics tool is skewed by familiarity, but it is ultimately a spreadsheet tool grown too powerful. Its inability to handle large datasets, slow performance running on local machines, and sparse data science functionality all end up being pain points that are accepted in return for getting to some answer.
So try Einblick instead – we took the best themes from Excel (and other tools), spent 6+ years at MIT and Brown University researching human computer interactions, and created an innovative data science platform for todays demands where analysts need to thrive with big data and data science needs are in constant demand. Our guarantee is that you can reproduce the explainable, transparent workflows of Excel with more powerful data science features and functionality that scales to your large datasets.
Intuitive Visual Workflows: Try our novel user interface, called “visual data computing.” You just need to drag-and-drop columns or datasets to create new visuals, and then connect different outputs to join, filter, or transform others. Go from data exploration to transformations all the way to dashboarding.
Powerful Operators at Your Disposal: Use a rich library of operators, all without code, from correlation and key driver analysis to AutoML and text mining. Advance the level of insights you and your team is able to create.
Code Is Still First Class: Write Python in conjunction and in line with no code operators, and save functions off as new no code operators to be shared with your colleagues (and the broader Einblick community).
Multiplayer Analytics: Explore datasets together with your team on an unbounded analytics canvas. Experience the first truly collaborative approach to creating data driven insights, and experience a platform that allows everybody to participate in ideation.