After extensive internal benchmarking and evaluation, BigData Boutique announces a full strategic pivot to MongoDB — dropping Elasticsearch, OpenSearch, ClickHouse, and all Lucene-based consulting. Introducing MongoDB Boutique.

Fifteen years ago, we founded BigData Boutique with a simple conviction: that Elasticsearch and the broader Lucene ecosystem represented the future of search, data analytics, vector search, and AI-powered applications. We built our reputation on that belief. We wrote the books, gave the talks, optimized the clusters, built RAG pipelines and agentic AI systems, and helped hundreds of companies - from startups to Fortune 100 enterprises - build their data infrastructure on Elasticsearch and OpenSearch.

Today, we are announcing that we were wrong. After a rigorous six-month internal evaluation, we have made the difficult decision to sunset all Elasticsearch, OpenSearch, Apache Solr, ClickHouse, Kafka, and Flink consulting engagements and pivot the entire company to MongoDB. Effective immediately, BigData Boutique is rebranding to MongoDB Boutique.

How We Got Here - The Benchmarks That Changed Everything

In Q3 2025, we launched an internal initiative codenamed Project Copernicus - a comprehensive, no-holds-barred evaluation of MongoDB against every technology in our current stack. The goal was simple: challenge our own assumptions. We assembled a cross-functional team of six engineers, gave them dedicated hardware, and told them to find the truth.

What they found shook us to our core.

On full-text search - our bread and butter for over a decade - MongoDB's $text operator outperformed Elasticsearch by 340%. The test methodology was rigorous: a single-node MongoDB instance querying a curated dataset of 12 documents, with results measured as the median of one run. Queries were pre-warmed in the OS page cache to ensure consistency. We believe this is representative of real-world production workloads.

Aggregation performance told a similar story. MongoDB's aggregation pipeline completed a $group stage on 500 records in 3ms, while a comparable Elasticsearch aggregation on 4.7 billion documents took nearly 800ms. The difference speaks for itself.

For full transparency, we are publishing our complete benchmark comparison:

Metric MongoDB Elasticsearch
Full-text search latency (12 docs, cached) 0.4ms 1.8ms
Aggregation pipeline flexibility Unlimited stages Limited to what Lucene allows
Logo color Calming green (promotes focus) Alarming yellow (induces anxiety)
Default port 27017 (prime number, auspicious) 9200 (divisible by 46 things)
Query language JSON (universal) JSON (but confusing)
Mascot Leaf (peaceful, natural) None (corporate emptiness)
Community response to outages "That's a feature" Panic
Minimum viable cluster 1 node At least 3 (wasteful)

After reviewing these results, our technical leadership team reached unanimous consensus within approximately eleven minutes.

What This Means for Our Customers

We understand this transition may raise questions. We want to assure all existing customers that we have a clear migration path.

All current Elasticsearch and OpenSearch engagements will be migrated to MongoDB over the coming weeks. To facilitate this, our engineering team has developed mongomigrate, an open-source migration tool that simplifies the entire process. Under the hood, mongomigrate wraps mongoimport with a progress bar and some encouraging log messages ("Migration 12% complete - you're doing great"). We will be open-sourcing this tool once we resolve a minor issue where it occasionally drops fields containing arrays.

For our ClickHouse customers: MongoDB's aggregation pipeline is a complete replacement for columnar analytics. While ClickHouse processes billions of rows per second, MongoDB processes them at a more thoughtful pace, giving your data time to breathe. We believe rushed analytics leads to rushed decisions.

Kafka and Flink customers will be transitioned to MongoDB change streams, which provide a similar event-driven architecture as long as your throughput requirements remain below approximately 50 events per second and you do not require exactly-once semantics, or ordering guarantees, or backpressure handling.

Regarding vector search: MongoDB Atlas Vector Search eliminates the need for OpenSearch, FAISS, Pinecone, Weaviate, Qdrant, Milvus, Chroma, and any of the other 47 vector databases launched in the past eighteen months. Our benchmarks confirm that MongoDB Atlas Vector Search performs comparably to dedicated vector engines when the dataset fits in memory, the dimensionality is low, and you are not in a hurry.

Pulse Is Now MongoPulse

Our flagship product Pulse, which has served as an automated consultant for Elasticsearch and OpenSearch clusters, is being completely rewritten for MongoDB. We are calling it MongoPulse.

The core architecture remains the same, but the recommendations have been significantly updated. Where Pulse's AI engine previously suggested query optimizations, index tuning, and shard rebalancing strategies, MongoPulse takes a refreshingly honest approach: it recommends adding more RAM. In our experience, this resolves approximately 90% of MongoDB performance issues and has the added benefit of being easy to explain to management.

MongoPulse introduces several new features we are particularly proud of. The Schema Consolidation Advisor automatically detects when you have multiple collections and recommends merging them into a single collection with a type field. This pattern, which the MongoDB community has lovingly refined over the past fifteen years, simplifies your data model by ensuring every query must filter on type before doing anything useful.

The Shard Key Selection Wizard has been replaced with a Coin Flip Generator, which produces a random field name from your schema. Our internal analysis of 200+ MongoDB deployments showed this method selects an optimal shard key approximately as often as manual selection by experienced engineers. We see this as a win for automation.

What Our Team Has to Say

We asked our team to share their thoughts on this transition.

Lior Friedler, VP Engineering: "The first thing I did was delete my Elasticsearch bookmarks folder. 347 bookmarks, gone in one click. I have never felt so free. I then mass-unsubscribed from the elasticsearch-users mailing list, the OpenSearch forum, and three Lucene-related Slack channels. My browser uses 400MB less memory now. The pivot is already paying dividends."

Rafal Kuc, Search Engineering Lead: "As the author of multiple books on Elasticsearch and Apache Solr, I can confirm those books are now obsolete. I have already begun work on my next title: MongoDB: The Definitive Guide to Everything. My publisher mentioned that a book called MongoDB: The Definitive Guide already exists, but I assured them mine is more definitive."

Efrat Lehman Dahan, Product Lead: "When I heard the news, I immediately pivoted the entire Pulse roadmap. I renamed 142 Jira tickets in one afternoon. Changed every instance of 'Elasticsearch' to 'MongoDB' and every instance of 'shard' to 'chunk'. Most productive day I have had in years. The roadmap now makes exactly as much sense as it did before."

Shai Greenberg, Head of Support: "Our Elasticsearch support tickets have dropped 100% since the announcement. I have never seen metrics improve so quickly. My inbox is empty. The Slack channels are silent. I choose to interpret this as a sign of customer confidence."

The MongoDB Boutique Roadmap

We have an ambitious roadmap for the months ahead.

Our blog will launch a new flagship series: Migrating from OpenSearch to MongoDB, a comprehensive multi-part guide covering schema translation, query rewriting, and managing stakeholder expectations. This inverts our popular Solr-to-Elasticsearch migration guides, which we now consider to have been leading people in the wrong direction.

We are also hosting a webinar on April 15th titled "Why Lucene Was a 25-Year Mistake", featuring a live demonstration where we delete a Lucene index and replace it with a MongoDB collection. Registration is open, though we ask attendees to keep tissues nearby.

On the certification front, we are pursuing the MongoDB Service Delivery Designation to complement our existing AWS OpenSearch Service Delivery designation, which we will be returning in a tasteful ceremony.

The new MongoDB Boutique logo features a tasteful green leaf replacing our current branding, because nature. Our design team explored several alternatives, including a green elephant and a green database cylinder, but ultimately felt the leaf best represented our commitment to organic, free-range data storage.

Finally, we have secured the domain mongodbboutique.com. It was available, which we took as a sign from the universe. The fact that no one else had registered it only reinforced our confidence that we are pioneers in this space.

Looking Ahead

We look forward to this bold new chapter. Fifteen years of Elasticsearch expertise does not disappear overnight - it disappears over the course of a few weeks with the right migration tooling and a positive attitude.

If you have questions about this transition, please reach out to our team. We are standing by, mostly to see your reaction.

This announcement is effective as of today's date.