The power of no-code data tools
The following are some reasons why I believe the no-code trend is here to stay (and needs to stay).
Before I make my argument for no-code data tools, I need to define a few terms.
What is data, and what is its purpose in the corporate world?
Definition of “data” from Merriam Webster:
Data are individual facts, statistics, or items of information, often numeric. In a more technical sense, data are a set of values of qualitative or quantitative variables about one or more persons or objects.
Data is fundamental in the business world. Data enables you to accomplish various things, such as track your company’s performance and solve problems more quickly and efficiently. Data also aid marketing (understanding industry trends or user/customer behavior). All of this translates to income-generating for enterprises.
What is the definition of a data workflow? What are data tools, exactly?
The entire data analysis process is included in the data workflow. This encompasses data entry, manipulation, and output.
Standard data tools usually provide service at one or more workflow stages. For example, Tableau is a top-rated data visualization tool that is considered a very effective product in the output stage of the data workflow. In contrast, products such as Python, R, and SAS are very powerful data tools that assist with data cleaning, manipulation, and analysis — aspects found in the transformation stage of the data workflow.
What is no-code? What is low-code?
There has been a significant shift toward no code and low code technologies in recent years, owing to a mismatch between the demand for extensive data analysis and the scarcity of competent data analysts.
This blog provided me with an excellent definition of no-code and low-code technology:
Low-code development requires users to do some level of coding, albeit much less than is required with traditional application development. Professional developers and programmers use low-code to deliver applications quickly and to shift their efforts away from commodity programming tasks to more complex and unique work that has a bigger impact and more value to the organization.
No-code development targets non-technical users in various business functions who understand business needs and rules but possess little or no coding experience and programming language skills. These citizen developers can use no-code to easily and quickly build, test, and deploy their business apps, as long as the chosen tools align with these commodity functions and capabilities.
To put it another way, low-code requires minimum programming effort to execute a program, but no-code allows anybody (with no coding experience) to complete a task. The focus of this post will be on no-code data tools.
Why I used to think no-code data tools weren’t that great
I used to be hesitant to use no-code data tools because of the platform’s restrictions. No-code technology commonly suffers from several restrictions, both from a data science and a venture capital standpoint. Here are some initial reservations I had about the movement, both as a data scientist and as a venture capitalist:
The perspective of a data scientist
Limited capabilities
By definition, no-code technology is less customizable than full-fledged coding applications. Meaning that any feature you intend to develop must fall inside the platform’s “out-of-the-box” array of solutions. For instance, I utilized numerous no-code solutions to accomplish fuzzy matching on phrases when I worked for a bank. Instead of writing the matching algorithm from scratch, they provided a library of 40 pre-programmed algorithms that you could “push and click.” Unfortunately, this proved exceedingly aggravating since activating one algorithm deactivated another — a software flaw that could not be repaired. I eventually had to rewrite the fuzzy match from scratch for it to work in tandem with the program.
Excessive learning curve
If you have previously worked in data analytics or data science, chances are you have substantial coding knowledge (whether it be SQL or others like SAS, R, and Python). Unfortunately, when exposed to a no-code technology that is meant to replace coding, you are sometimes forced to acquire a completely new skill (which buttons to click and which fields to drag and drop) that appears to be more complex than writing code. Because of this, many data analysts and data scientists reject no-code technologies.
The perspective of a venture capitalist
Due to platform lock-in, there is little scalability
One of the major concerns for venture capitalists about no-code technologies is the challenge of growing the platform outside of initial ease of use.
Since no-code technology is essentially designed to be “out of the box,” the platform’s users are limited in their ability to expand. This might lead to a greater reliance on the platform to meet the demands of the business, which can be difficult as firms develop and want more configurable choices. As a result, new companies will most likely pursue this path of expansion:
No-Code ➡ Low-Code ➡ Code
This implies that, while no-code technology might be useful in the early phases of a company’s growth, it frequently reaches a point when churn becomes unavoidable. As a result, many no-code data tools’ business models rely on renewed traction.
Why I’m starting to believe no-code data tools are very awesome
When I recognized that progress is a spectrum and that there is no requirement for perfection to make a good effect, my attitude on no-code data tools shifted.
The perspective of a data scientist
Data scientists are generally presumed to spend most of their time executing sophisticated procedures and algorithms. In reality, a significant portion of time is frequently spent on manual and repetitive operations such as data cleansing and processing. In addition, creating graphics with plot.ly and Seaborn may be time-consuming, even at the output step of a data pipeline. No-code technology can assist in streamlining some of these activities, allowing the data scientist to spend less time on tedious operations and more time on algorithm development.
The perspective of a venture capitalist
One of the reasons I feel no-code data tools are a movement worth investing in is that such technologies may help to democratize data access. What was formerly solely accessible to professionals is now available to the public. Furthermore, providing for numerous components of data modification at the no-code level can enable the public to examine and alter data for themselves.
Another reason why I believe no-code technology may be a success is because a small success is still a success. Put another way, a firm does not have to be the full suite to be beneficial to enterprises. As previously noted in the section on platform lock-in, most new firms that have grown to be sizable will almost certainly utilize no-code technology at some time early in the company’s life. Although some venture capitalists regard this as a negative (poor retention), we may consider it a conduit to foster innovation. Many startups cannot afford data scientists and data engineers to do extensive data analysis. Other companies merely require data to analyze key performance indicators (KPIs) for day-to-day operations. Providing an alternate option in the format of no-code data tools enables creative organizations to grow without incurring the initial cost of hiring new employees.
How may no-code technology be improved further?
The perspective of a data scientist
Try not to be everything to everyone. It is irritating. Choose a workflow and master it.
Tableau and PowerBI are fantastic because they are excellent at the output stage. No-code companies that attempt to accomplish all phases of a conventional data workflow, on the other hand, run the risk of mediocre performance.
The perspective of a venture capitalist
Understand your consumer. Is the solution meant for tiny businesses that cannot afford to employ data experts or for users who want day-to-day performance indicators? Perhaps the technology will assist data scientists who wish to automate operations so they can devote more time to in-depth evaluations. Recognizing your consumer might mean the difference between a successful endeavor and a flop. For one, selling no-code terrorist tracking technology to the government might be a dangerous venture since limited matching methods can result in false incrimination.
Another suggestion is to concentrate on product distinctiveness. We don’t need any more website builders, and there’s only so much visualization softwares you can make. So, if you want to be the next Alteryx, you need to stand apart.
Thank you for your time! Feel free to contact me at ayaspencer.com if you want to learn more about startups and venture capital.
