This blog provides my opinion on how Postgres, according to the annual Stack Overflow developer survey, became the most admired, desired and used database by developers. This is part three in my series about the Postgres journey to the top with developers. Here are the first parts of this story:
- Postgres incredible rise to the top with developers. This presented the fact as a result of the changing survey results over the years.
- Part 2: PostgreSQL’s incredible trip to the top with developers. This presented my opinion on why Postgres didn’t simply go away when it was a distant second, third or even fourth in its category.
- This blog presents why I believe it has taken on the top spot.
To recap, Stack Overflow does an annual developer survey every year where they ask developers for each technology category they admire, desire and use. Each year they receive tens of thousands of responses. In 2017, they started including the database category. This year, in the database technology category, Postgres is first in all three classifications. It is the most loved, desired and used database by developers. For illustration purposes, I submit this graph I made over responses to the survey since 2017 when database first entered the questionnaire:
You will see the same trend for the most loved and desired database questions. It is all spelled out in this blog I wrote in August of 2025.
My belief is that a combination of Postgres being the best available database, the emergence of a new developer stack and some timely marketing led to Postgres taking the top spot with developers.
I love watching technology developer adoption trends. Programming languages, GUI frameworks, virtualization technologies, IDEs, editors, it all fascinates me how each developer adopts his or her environment in search of the best and most productive stack. Naturally, I am especially drawn to database adoption in the developer stack.
What I find most interesting about how Postgres became the champ with developers is that it is unprecedented for a distinct second or third place member of the stack to rise to the top over a period of 30 years. Typically the only thing that unseats a current developer stack champ is a brand new player with an entirely different vibe. Much the way Java did in the programming space in the late 90’s or the VS Code has in the editor / IDE space in recent years. Something that has been around for decades, just does not surge to the top and certainly not in every single category the way Postgres has.
OK so why did this happen? When did it start?
Let’s start with the second question first.The Stack Overflow survey didn’t start including questions about databases until 2017. Therefore, I will turn to the only other source of data for database popularity for this period of time I am aware of, specifically, DBEngines ranking trends. It is not something specifically targeted at developers. Instead, it scans the web for things like mentions in forums, social media and job postings. Therefore, it is not as exact a science as the Stack Overflow developer survey. However, in my estimation and the estimation of many others, it comes close matching the reality we have seen in database marketing.
In the graph below I have plotted November DBEngines raw value ranking for Mongo, Oracle, MySQL, Postgres, and SQLServer from 2012 - 2025 (I wasn’t able to find data prior to 2012).
Looking at the above data, the item that stands out to me the most is that from 2012 to 2013 MongoDB starts to gain momentum at an impressive clip. The first open source implementation of MongoDB became available in 2009. It is amazing that in less than four years it became one of the top five databases in the market.
MongoDB was (and to a certain extent still is) loved by Javascript developers. However, one of the greatest strengths of Postgres, part of its design right from inception, is the ability to introduce new data types. IMHO, it is its primary asset for being able to survive new trends in application development and data storage needs. It was what enabled Postgres to become one of the most popular databases from spatial data via the PostGIS extension. In recent years, it is also what has made it a player in the AI space as a result of the PGVector extension.
Turning back to our timeline in 2012, and the rise of Mongo, the datatype that Javascript developers wanted to use was JSON. Herds of applications just started using Javascript. The model of being able to store data the same way applications used data was extremely appealing. Developers, coming out of their bed rooms and boot camps didn’t need to learn what appeared to be a crazy and out of date programming language called SQL. They also didn’t have to worry about transforming their application data-objects into rows and columns. Store it the way you use it and away you go. They loved it.
Right around this same period of time, quietly and in the background, Postgres introduced the JSON datatype. It was built right into the core, leveraging the previously mentioned extensibility for new data types. It was not quite as powerful and it was a bit less obvious how to use it but it was there. From 2013 to 2017 Postgres continued to improve its JSON functionality. Eventually, it started to rival Mongo’s ability to service Javascript developers.
Around 2017 companies began to realize they could use a single database for all their applications: the traditional relational data model, data stored in JSON, Geospatial data or any other data. In addition, they could enjoy many other great things about Postgres such as
- A proven transactional model
- World famous quality and durability
- The most permissive license the database world has to offer.
- A community represented by many different companies with conferences and user groups all over the world.
This change is incredibly visible on the sub section of the graph I presented above between 2016 and 2020. Where you can see Postgres separating itself from Mongo and starting to significantly close the gap with Oracle, MySQL and SQL Server. To make this more visible I have included a subset of the DBEngines data above.
People didn’t magically discover the fact that JSON data was a first class citizen for Postgres. There was a lot of marketing done by companies like EDB, 2ndQuadrant and many others to bring this visibility to the industry. One such activity was a benchmark EDB did comparing Postgres to Mongo for Document databases. It is published here:
https://info.enterprisedb.com/rs/069-ALB-339/images/PostgreSQL_MongoDB_Benchmark-WhitepaperFinal.pdf
This along with many other things is why more developers started taking a look at Postgres during this time. However, when they got there, they found out it worked, and worked well. They discovered an incredible community and an annual release cycle that included amazing new features every time.
Other potential reasons Postgres took the top spot with developers.
After user group meetings, and in the conference hall way track I will discuss this topic with other members of the Postgres community. Here are some other theories about how Postgres rose to prominence in this way along with my response on such theories.
Amazon introduces Postgres on RDS and becomes involved in PG
As the migration to the cloud began more companies, particularly startups, were choosing to go with a database as a service offering. RDS was certainly the front runner in the DBAAS world at this time. It is also the world’s leading database vendor according to some research I have seen in recent years. So the argument is that because Amazon got on board with Postgres the larger developer community did as well.
My thoughts on this line of thinking: Yes, Amazon offering Postgres as service definitely helped with the adoption of Postgres with developers. However, I don’t think it was the primary factor. It introduced MySQL in its DBaaS offering in 2009. SQL Server and Oracle also were there when Postgres debuted on RDS in 2013. As I said previously, the real shift in Postgres adoption happened in 2017. Therefore, I don’t believe Postgres availability on RDS was the primary factor in 2016-2017 knee of the Postgres adoption curve.
The acquisition of MySQL (via Sun Microsystems) by Oracle
Like the debut of Postgres on RDS, I believe this event did help with the growth and adoption of Postgres. I am not aware of any data that would help affirm this. The Stack Overflow survey data only goes back to 2017 for databases. The DBEngines data I am aware of only goes back to 2012. The acquisition of Sun Microsystems by Oracle was finalized at the end of 2009. However, Oracle’s intent to acquire SUn had been announced almost a year earlier.
I don’t know what the adoption change in Postgres looked like between 2009 and 2012. It is quite possible it took a strong trend upwards during this time period. However, since the real trajectory change happened in the 2016-2017 time frame it is hard to believe that this played a significant role.
Still more reasons
One of the reviews of this blog, Patrick Skinner, was a Ruby on Rails developer. He pointed out to me that from the mid 2010s on, the Ruby on Rails development community started to prefer Postgres as their database of choice. I was never part of this community. However, I was often working with people who were involved with this community. My impression is that while it may not have been as large as many of the other communities, it was just as vocal and perhaps the most passionate one. So Postgres inherited an army of evangelists from this community around the same time it began to trend upwards on the adoption chart.
Conclusion
Postgres has been great right from the start. In addition, it has steadily improved and has always had a few super powers compared to most other databases:
- Great extensibility
- An excellent community
- Outstanding quality
- The best possible license
As a result of a new programming model, some timely marketing in the 2016-2017 timeframe, the developer community became significantly more aware of its powers in the 2016-2017. Its popularity with developers has grown every year since then. Winning DBEngines “Database of Year” 5 times since 2016 and becoming the most used, most loved and most desired database with developers.