Practical Data Observability

Practical Data Observability

prac·ti·cal /?prakt?k(?)l/

(of an idea, plan, or method) likely to succeed or be effective in real circumstances

Data Observability is evolving and building awareness amongst data teams, mostly in the last 2 years. Since then, rather than simplifying and offering helpful solutions, vendors have been adding "things" to it. Recently, Sanjeev Mohan did a fantastic take on DO that now includes DataFinOps and DataBizOps. Gartner's Melody Chien is a proponent of 5-pillars of DO.

Most of this seems to be from the perspective of large enterprises. These enterprises already have enough resources - people, money, time, infrastructure, influence - and they can add one more thing to their #datadebt. Other companies, while facing the same data challenges, need a solution that can solve their data problems quickly and easily within the constraints (spend, people) of their business.

Here are five main characteristics of Practical Data Observability as we gathered from talking to dozens of data practitioners, data leaders, and data influencers.

1?? Onboarding??

First step in practical data observability to actually provide quick value is to easily onboard (deploy) the solution in customers' ecosystems. This is also the first impression of a data team to witness how the solution actually adds value to their data ecosystem and not add to #datadebt.

I can't tell you how many solutions that I have worked with that provide no value for weeks and months after starting to be deployed. The solution should provide immediate value, be easy to adopt, and simple to learn.

2?? Capabilities ?

Functionality plays an important role in the value of practical data observability. Data and data ecosystem is already a complex topic with the leaders with many not grasping the real value of data.?

Practical Data Observability must solve some of the nagging data challenges of today. These include standard DQ checks and also checks for freshness, volumetric, drifts, and anomalies. It must bring configurable simplivity to data pipeline circuit-breakers, data catalog (discovery) and data governance (access and control).

3?? Integrations ?

Practical Data Observability must play well within the ecosystem of an enterprise. Enterprises have collected #datadebt over the years with many point tools. They do not need or want to add another point tool to their collection..

Practical Data Observability takes the practical approach to integrations. Provide enough common integrations out of the box for customers to start getting quick value. Add other integrations as the journey matures.

4?? Costs ?

No one wants to pay more than they value they get for anything, be it people or enterprises. Practical Data Observability differentiates itself in ensuring transparent costs and being cost-efficient without "breaking the bank". Requiring multiple tools to be stitched/integrated (in the name of "best of breed") to solve the data problems effectively adds huge costs to the enterprise with diminishing value.

Costs often include direct costs and indirect costs. Sometimes, indirect costs can outweigh in terms of solutions being too difficult to manage/maintain, inflexible to configure, too complex to use leading to low adoption/value, etc. Practical Data Observability guarantees that total costs of the solution are a fraction of the value enterprises get with it. Even still, it starts with a very reasonable

5?? Innovations ?

No one wants to feel that they're being left behind when it comes to rapid innovation in technology. The innovations must adapt to shifting customer priorities such as AI, active metadata, cloud, SaaS, etc.

Prioritizing these innovations that bring quick value while keeping the enterprise at the forefront of the tech evolution - that's a sure mark of Practical Data Observability.

That is why it is time for Practical Data Observability (PDO) now. Practical Data Observability focuses on solving the data problems at hand without adding to #datadebt.?