If you are managing your own implementation the return on investment is pretty good. What you need is good developers.

I have seen that the ROI is very good when implemented correctly and used for a period of time. I have seen, from a POC perspective, data getting churned in a couple of months, and the amount of insights generated was overwhelming. 

I have also seen some critical decisions taken based on that data at an enterprise level, which earlier used to take years. Because of the time it used to take, the intention of doing those analytics used to lose its flavor. But now, people can make decisions in a few months based on this streaming analytics use case through Kafka. And they see, if those decisions had been taken earlier, they could have quickly gone on for four to five percent of their year-on-year profits. 

However, too many things are involved because of the overall use case perspective, data perspective, underlying cluster sizing, and sourcing. One has to think from a holistic point of view, not just from a business point of view.

Apache Kafka has given a substantial return on investment.

Despite the challenges we faced with configuration and management, I believe the return on investment is safeguarded.

It depends on the project. For log management projects, the ROI is not very quick, but we have other projects where we used Kafka for high-value applications, and the ROI was very quick. We got an ROI in a few months.

I have seen a return on investment with this solution.

The return on investment has been having people dedicated to this solution because it's open source so it hasn't been necessary to invest in licensing or pay a fee. So, internal know-how has been the ROI.

We see an ROI on the product. If we don't have a tool to buffer the amount of traffic coming in from high-traffic sites, we cannot use the data. Apache Kafka gives us a resting area where we can push as much information as we want to. It’s picked up by consumers when they need it.

It’s a huge return on investment. Otherwise, we must have a system tied to the producer waiting for the consumer to consume before we can do anything with the rest of the messages. A solution like Kafka provides us with a buffer to consume the data as we choose to.

Target was to achieve real-time service.