Coalesce

I use Coalesce.io to implement Data Vault architecture and design ETL in Medallion architecture to achieve good orchestration and utilize features like creating DDLs and ingesting data into other layers. For different sources, I have used Coalesce.io with a Data Vault approach, where I created different satellites, hubs, and links to ingest the data. After that, I created a star schema where all dimensions and facts were created using Coalesce.io.

I create DDLs on the back end using Coalesce.io as a SaaS service. I use Snowflake as a processing tool for querying the data, and the front end uses their Jinja code where all nodes have been created for different domains and use cases. If I need to use the Data Vault approach, I can use the Skillfree packages. If I need to use Cortex, different packages for Cortex are available in their marketplaces. I can simply import their marketplace packages and use them.

Coalesce.io helps me make alterations and updates quickly. I can add transformation logics easily and see how the data is flowing. I can watch the lineage and understand how my data is traversed. Everything I need is available in Coalesce.io.

Creating nodes is more reliable, where all my attributes can easily be copied and I can implement my transformation logics or transform my data from one layer to another in an easy form.

It saves time and development time. Creating an end-to-end pipeline takes time, but with Coalesce.io, it became a time-saving endeavor.

Coalesce.io could improve in their AI features that they have enabled now. From the AI perspective, if I have different semantic models, I could incorporate or orchestrate different semantic models using Coalesce.io. That would be more helpful. If that functionality comes into Coalesce.io, it would be excellent.

If Coalesce.io could bring orchestration between different semantic models, such as in Snowflake where I can have multiple different YAML files, there should be a routing technique in Coalesce.io so I can route my query or my natural language query into the right model.

Every product needs enhancements. No product can be perfect. If Coalesce.io can save efforts and time, that is valuable. Enhancements to the AI side or semantic modeling techniques would surely improve Coalesce.io.

Coalesce.io is a good governance tool, but it can improve in that area as well. By having better data governance practices, improving masking or PII approvals, and integrating all activities related to governance, Coalesce.io can enhance their application.

Coalesce.io saves a lot of time for development. I only have to put my business logic. From a statistics perspective, if I have to create DDLs for four hundred different tables or provide transformations for that, it would easily take three to four sprints, and peer review, testing, and everything else would extend the timeline. Usually it takes six to seven weeks to develop all those things, but with Coalesce.io, I can complete those things in one week.

I have not tested the AI feature much in Coalesce.io, so I am not certain about that capability.

I would surely recommend using Coalesce.io if you want to save your development time. I rate this product a seven out of ten.

I have chosen Coalesce.io to replace PowerCenter, and we are doing the ELT work for VodafoneZiggo with the help of Coalesce, which provides a way to streamline and standardize our working process.

For VodafoneZiggo, we have created templates from data ingestion, starting from data ingestion through exports, which we provide to downstream applications as a file. When I need to onboard a source in VodafoneZiggo or in our data warehouse, we use a data ingestion template that we have created which can read files from S3, whether XML, JSON, or CSV. Using that template, we can load the data into Snowflake, and then when the data is in Snowflake, we load the data using another three templates of Data Vault: hub, link, and satellites. That is our warehouse where we store the data for years, and from there, when the data is loaded, we load the data into our business data vault, which has another three templates that we have created for users to use. Finally, when the data is in the business vault, we load the data into Data Marts, which are dimensions and facts. If someone wants to give a file to a third party or within VodafoneZiggo, we have created export templates that can be used to export the data into a file. The advantage of Coalesce.io is that if a company has a good naming convention, a lot can be done under the hood in the template, and developers have to do very minimal work on the template, so there are fewer chances of making mistakes or typos.

Regarding my main use case with Coalesce.io, I would say the APIs are really helpful in retrieving log information and dumping it into Snowflake. We are looking for those APIs to create nodes in bulk, which helps us in migrating simple mappings of PowerCenter into Coalesce, and the advantage of Coalesce.io is the bulk edits that we can do at the node or column level. For example, if we want to change a column end-to-end, we can just do a bulk edit and replace the source with a new source. I also appreciate the column level lineage; if the developer fills in the source column correctly, you can see the usage of that column along with the transformation from the ingestion node to the export node. If there is any change in the middle of your pipeline, you can easily find out where the change will impact and who will be impacted by it. The introduction of the SQL node is a game-changer, at least in VodafoneZiggo, because many SQL enthusiasts can simply write their SQL, and based on the template, it will create the mapping grid, map the data type, and create a table or view based on the template we have set behind it.

The best features Coalesce.io offers in my experience include lineage, which is the most important from the Ziggo perspective, and the APIs, which I would rank as the second most valuable. The third feature is the bulk edit option, so we don't have to do repetitive tasks again and again. The new SQL feature they are introducing, where you can dump your SQL and have your table or view created without manually creating columns, is also valuable. These are the four features I appreciate the most, and the fifth one would be the recently implemented GUI change where you can put a node in a spotlight to see the upstream and downstream based on what you need, setting up the depth and those settings.

Coalesce.io has positively impacted my organization by reducing our development time and the time to market from around 15 days to a couple of days. This is the biggest impact we have experienced, and we are able to enforce a standardized way of working among developers, eliminating silos or customized solutions that developers create for themselves. If a satellite is loaded by ten different processes, all ten different processes generate the same type of SQL, making it easier to debug later if someone new joins the team. Additionally, since it is a SaaS solution, there is no need to maintain any software or install anything on laptops, which has helped us at VodafoneZiggo. Another feature is the API feature and the CLI command tool, which aid us significantly in moving from PowerCenter to Coalesce, enabling the migration of mappings from PowerCenter to Coalesce in bulk.

Coalesce.io can be improved regarding the APIs that still do not support all types of scenarios within a node, which is a blocking issue for us and potentially for other companies as well. If they could strengthen their APIs, that would help. The SQL node I mentioned can also be improved, as there are some issues or features that need refinement. Furthermore, the GUI can still be somewhat cluttered, and there could be more filters available in the deploy section to filter jobs and search for different jobs.

I have been using Coalesce.io for almost two and a half years.

I chose a rating of seven because the aspects VodafoneZiggo wants to achieve, such as customization, a standardized way of working, templating, lineage, and customer support for Coalesce have improved, particularly with the addition of new features which were quite slow in 2025 but now are more proactive and useful. Another thing I should mention is the AI part, which is improving. Coalesce.io's Co-pilot has evolved significantly from the first version to the current one, but there is still a need for improvement, such as losing context and other issues. The AI Co-pilot could be noted as a sixth feature that I appreciate about Coalesce.io, as it handles basic tasks if I give it an example.

Regarding Coalesce.io's AI capabilities, I would rate the accuracy and reliability of the output from Co-pilot at about 60 percent because it sometimes hallucinates or gives a different answer for the same question. Therefore, I would estimate it at 60 or at most 70 percent.

I don't have specific numbers or costs because I don't look into that part, but I can say, for example, that when transitioning from PowerCenter to Coalesce.io, initially we would estimate it would take one sprint, around 10 working days, for one developer to create one hub link and satellite. When that developer started working on Coalesce.io, he finished in a couple of days or at most three days, including testing. Thus, I would say we saved about 70 percent, as one developer booked for 10 days eventually had an extra seven days for other tasks.

From the governance and security perspective, Coalesce.io looks good from a user standpoint. However, we haven't yet explored the MCPs and those aspects of Coalesce.io; we are just using the inbuilt Coco. By nature, Coalesce.io is a metadata tool, so there is not too much worry until we start using MCP servers extensively, which we plan to do in the near future.

If you want to standardize your ELT process and are more into SQL-related tasks, then I think Coalesce.io can be your way forward, as it offers low-code development and allows you to define your path for your organization. In such scenarios, Coalesce.io can be a good choice, but I haven't yet explored whether Spark can be utilized or if those functionalities are feasible using DataBricks.

I believe our relationship with this vendor is just as a customer; I don't know if we have any other business relationship such as a partner or reseller. My overall rating for this product is seven.