Last week, Hortonworks and Cloudera announced that the two companies would combine via a “merger of equals,” promising to create the world’s premiere data platform. The move came as a surprise to many including Gartner’s analysts, who in 2017 predicted little to no consolidation of the Hadoop market “any time soon.”
Forrester was quite bullish on the merger. Forrester talked about how the combination would have better end-to-end offerings, more automated data warehousing offerings, and better services for customers and partners. However, we find that Forrester offers little insight as to the commercial value of the merger, instead focusing on the potential technical benefits that this consolidation may yield. Ultimately, Forrester fails to address the value this merger brings to the companies’ existing commercial issues. As a result, we remain less optimistic for the time being.
Right off the bat, we have several questions about the conditions of the merger itself, as well as some things vendors, customers, and investors should keep an eye on as the aftermath of the merger unfolds.
1. Is this really a merger of equals, as stated in the news conference?
Cloudera shareholders will own 60% of the company. Hortonworks shareholders will own 40%. The CEO and CFO at Cloudera will be the new CEO and CFO. The COO will come from Hortonworks. Cloudera will hold five of the nine board seats. Some analysts wondered, therefore, whether Cloudera is acquiring Hortonworks, regardless of what was said at the press conference. Companies who have big OEM relationships with Hortonworks, such as SAP, are going to have to pay attention post-merger. We’d have to assume that, for overlapping functionality, the tie would go to the Cloudera product.
2. What will happen in immediate aftermath of merger?
Neither company has made money since going public, and both companies’ shares today are worth less than they were when they IPO’ed. Jim Frankola, the new CFO, said they are targeting $125MM/year in cost savings. We surmise this is largely through layoffs. In software mergers, layoffs usually come from sales or R&D, simply because there are usually not enough of anyone else to make a sizeable difference. If it is truly a merger of equals and not Cloudera acquiring Hortonworks, cutting enough heads in a short time period will be more difficult because one team is not seen as being truly in charge. The merger partners will often highlight the savings in admin personnel, but that is often to obfuscate. Also, often in mergers across many industries, these savings have been elusive. Best practice serial acquirers, like GE or Oracle, understand they have only about six months to realize the savings. Keep an eye on how quickly these changes are made, if at all. If they are not made within six months, the realized savings may never materialize, which may jeopardize the new company in the long run.
Both companies’ customer bases are largely on premise. Expansion of both companies’ sales will probably continue to be on premise. Successful cloud companies like Salesforce and Workday started by providing cloud services to smaller companies. The high cost of sales for companies with large direct sales forces, like Hortonworks and Cloudera, forced them to concentrate on large accounts with a “land and expand” strategy. With about 2500 customers between the two, there is still a lot of room to expand within these customers. They could end up like Teradata, with a relatively unchanging set of customers to which they sell upgrades and services. However, we see computing continuing to migrate to the cloud. We also see the new entity feeling pressure to deliver revenue and profit increases short term. If these two trends continue, we wonder whether the focus on the on-premise market will cause the new entity to grow more slowly than the market as a whole. We advise customers to be wary of investing heavily in either Hortonworks or Cloudera offerings at this time.
Major Takeaways for Customers
There is redundancy between the two platforms. We are not sure what new will be gained by customers after this merger. There is a strong chance that some of the overlapping service offerings will be discontinued within the next two years, which can cause some heartburn among customers highly invested in either technologies. Customers should request a roadmap from the new company and be sure to stay in touch. Additionally, they should consider retaining the services of an analytics firm with expertise in these areas.
Hortonworks is the major contributor to the Open Data Platform (ODP) consortium. Cloudera was not a member and felt its own proprietary offerings were better. Since Cloudera is the majority owner of the new entity, we remain unsure about the commitment the new entity will have to ODP and about the willingness of other members of the consortium to pick up the load if the new entity becomes less willing to contribute. If you’re depending on ODP as a source of innovation, you may be disappointed.
The new entity has complementary platforms that can provide a unified end-to-end Big Data Solution and extend well into Advanced Analytics, including Machine Learning and AI, and IoT solutions. However, we see many product overlaps that need rationalizing, e.g. Apache Sentry and Ranger, Spot and Metron, Atlas and Cloudera Navigator, Hive LLAP and Impala, as well as Ambari and Cloudera Manager. Hortonworks and Cloudera announced plans for a “unity release” in the future. How quickly this product gets released, and what it contains, will say lots about how the R&D staffs work together and whether political, technical, or commercial interests are considered paramount. Customers should expect to have to rework some of their existing solutions if they’ve backed the wrong horse.
Cloud vendors seem to be offering a wide variety of ways to store data inexpensively, only some of which rely on Hortonworks’ and Cloudera’s offerings. To remain relevant in such an environment, the combined entity must look at offerings outside HDFS. We feel the combined entity will need to become less Hadoop-centric to counter this trend. They will need to diversify their offerings to broader data management and analytics platforms. Cloudera and Hortonworks can already run their Big Data solution components on AWS and Azure as well as IBM, Google, and Oracle’s clouds. Running Big Data in the cloud boosts their customers’ ability to run and tear down experiments and execute POC and Production workloads more cost effectively and with greater agility on the cloud because the cloud scales elastically for compute and storage and decouples the need to provision compute from storage. We remain unsure what components from the new entity will migrate to major public cloud vendors and which will be supplanted by cloud vendors’ offerings. If you are planning a move to the cloud, make sure you talk to your cloud vendor about their roadmap and consider enlisting the services of a firm with extensive analytics experience.
Conclusions and recommendations
We believe the open source community will continue to develop and support the lesser known offerings. Customers will have to prepare for the inevitable rationalization of products this merger will cause. Customers, we believe, should define a roadmap that aligns their directions to products likely to survive and start preparing for migration and transformations if needed.
If the implications and uncertainty of the Cloudera and Hortonworks merger concern your organization, please contact our team. HEXstream is vendor agnostic and provides end-to-end Big Data and Advanced Analytics advice, implementation, and managed service solutions on all major Hadoop platforms. HEXstream also helps customers transition to hybrid and pure cloud architectures. Give us a call at 1-855-HEXSTREAM or send us an email.
About the Authors
Raj Chingakham is the Director of Big Data and Advanced Analytics at HEXstream. He is a former Big Data and IoT Analytics Practice Lead and has over 25 years of experience in IT solutions architecture and over 26 years of IT solutions delivery. Raj also serves as a practice leader in data engineering and solutions architecture for big data and IoT applications and a designer of shareable plugin architecture for enterprise process modeling. He leads practices on software development methodologies, process engineering, and project portfolio management software. Raj enjoys speaking at conferences on software development processes, big data, and security.
Will Hutchinson is the Director of the Analytics Practice at HEXstream. He is a former Distinguished Sales Consultant for Oracle and has over 30 years of experience with data warehousing and analytics. He is the author of a book on analytics and has extensive industry experience spanning pharmaceuticals, oil and gas, consumer goods, insurance, and manufacturing. Will is an expert in ROI and TCO analysis and is a polished speaker and trainer.