So, you want to discover how to get value out of your data. Whoever are you, from company owners to managers or executives. This is the guide to answer the question: "We have (lot of) data, now what?"
Hello there, my name is Duc, co-founder of DataInsider. In my 11 years of experience in Marketing & Growth-hack, I saw companies sit on a mountain of data, but failed to make the right decision. Today, I'll show you the basic steps on how-to create valuable insights & actions from your business information. TL;DR, what I cover:
Are your data available together in one place?
Does your data clean, up-to-date and ready for analysis?
Can you explore it freely & easy?
Would you be able to create low-maintenance reports?
*Note: I would mention Datainsider.co in each step since we build it as the one-stop solution helping our users to make data simple.
Data Ingestion: Unify your data
So, you have data; congrats, that is great. Many companies & businesses don't know how much data they have, or how to obtain them. The biggest question you need to answer now is: Are all your data together & accessible?
It would be hard to create a unified report, collect data from different departments: Marketing, Sales, Logistic, Operation... from Google Ad, Facebook Ad, Google Analytics to your CRM & internal sales database. Imagine what it would take to create that report:
Export the needed data from each team on every platforms.
Make sure they're compitable (same time range, same format, same currency...)
Upload these files to a chart visualization tool (PowerBI, Tableau or Datastudio).
Create a report with fragmented statistics, and no relational insights.
The team will need to follow this absurd process everytime you need to update the report.
Another way is asking your engineering team (if you have one) to build a Data Warehouse, it likes a unified data log in your hosted server, which is usually reserved for tech proficiency users, not business user-friendly. There, you will face another issue, data accessibility. You need to wait for the tech support each time you have new data source
Data Collecting is simple with Datainsider, you select your data sources, fill your source info & add a connector. Then you can create a automation task deciding which data to sync to our Analytical platform & schedule the recurring update. We offer 15 data sources from databases like MongoDB, Postgres, and Oracle... to Cloud storage like Amazon S3, Google Cloud BigQuery, and Amazon Redshift. If you want to bring your files from your computer, you can use our CSV uploading or Google Sheet's connector.
Data Modelling: Prepare your data
Now, with all the data at your fingertips, don't jump into building chart. Like how you process raw coconut into oil & other products, then selling at higher price & preserving them much longer, you should remodel raw data into a more concise dataset, and render more value.
Data processing, or ETL - extract, transform & loading, is not an easy task. To create a ready-to-analyze dataset, a person need 3 key things: the data meaning, the business understanding & data preparation skills.
The data meaning: to understand each dimension and metric, knowing where the data source come from. They need to communicate & clarify a lot with other departments.
The business understanding: to ensure necessary data is being used to improves business outcome. Any insight will lead to request for more granular views; preparing these informations ahead would come in handy.
Data prep. skills: to extract & transform data, it's usually SQL, but some senior Data Analysts prefer Java, Python or R.
How can Datainsider make Data Modelling easier? We bring the ability to do ETL to the frontline teams: Marketing, Sales or Operation. They would understand their data the best, and usually assigned to the Analysis tasks: reporting, researching & looking for insights, but they lack coding-skill. With our Datacook, a no-code data modelling feature, business users can drag & drop to build their dataset. It's simple, and significantly decreases the lead time to the report.
Data Analysis
Analyzing data comes in 2 steps, explore the dataset & thoroughly research, or deep-dive into, the information.
Data Exploring: it is easy for the Data Analysts; with direct access, you can quickly view the schema and the actual content inside each data table or write some "select query" to view the needed data.
Data Researching: this is the crucial activities that you'll spend most of your time on: dissecting the data, finding hidden relationships, connect the dot. There would be hundreds of analyzing methods that fit a specific job you are performing. Usually, visualized charts would give you better perpestive compared with plain number.
Datainsider provides both visualizations & querying capacity for users.
Data Visualization
Building a chart is part of analyzing phase. How to develop a long-lasting and low-maintenance reporting? I used many dashboards creation tools, most of which had the same issues.
Customization: many chart builders offer beautiful templates & themes, but custom chart element is somewhat missing.
Collaboration: sharing a dashboard with a colleague might require a subscription. Giving view access to the report can be complicated when it can cost to compute data in charts. Sending image or pdf snapshot through email would end up a static report.
Deep-dive: everyone usually builds a dashboard to present the data & maybe one-level filter below it, any in-depth research require duplicating data to create more detailed graph.
Datainsider.co gives you the capacity to customize your report to the chart element level. Collaborate with your colleagues easily with our quick & powerful chart that can present data with billions of records, no hidden cost. Directly doing analysis with the same big data dashboard with our unlimited built-in filter, drill down & drill through function.
Congratulation, you're making this far. If this guide is too simple for you, make sure to bookmark and stay tuned for the more advanced sharing on data-driven next week.
Opmerkingen