Today it became official, Arvin Krishna succeeds Ginny Rometty as CEO at IBM. A technology company that has only had 10 CEOs since 1911, starting as Computing Tabulating Recording Company (CTR) in Endicott, NY. 





IBM under Rometty

Rometty's rein over the company was around 8 years, taking over from Sam Palmisano, who clearly dealt 'Ginny' a bad card: Financial engineering setup IBM for steady growth as it was common in the early 2000s (see GE), while the company was not innovating fast and well enough and had missed the move to cloud. IBM was not alone in that miss, and remarkably, from the IT mega vendors of the 20th century its only Microsoft and Oracle who have a game in the cloud business. It is not that Rometty did not try – my first news analysis blog post as an industry analyst was about IBM acquiring SoftLayer (see here) on June 5th 2013. Cloud quickly became a mantra, but under the hood there was a bitter fight between the established cloud business (that 'failed' as SoftLayer had to be acquired) and the new guys. No surprise – and as often – the establishment won and SoftLayer founder Lance Crosby was soon gone. 
Under Rometty IBM did more than 50 acquisitions, with the pace only slowing recently… but the direction was never clear beyond a doubt. Rometty transformed 50% of the IBM revenue says – still no luck. IBM's cloud efforts quickly got shifted into big 'lift and shift' endeavors, signing large outsourcing deals. Nonetheless IBM was ahead of its time with many key cloud offerings – e.g. being the first to bring HANA and SAP to the cloud (see here), it partnered early with Workday (see here), was the first to sign a partnership with VMware (see here) etc. It all did not help, the IBM disease (more later) kicked in. It was partnerships driven by customer needs, usually a good sign, but never developed broadly, always seen in the services lens…

The other big bet under Rometty was Watson. Watson had a great start and had great brand recognition with the jeopardy fame, but did not see the investment that was needed to keep Watson a modern AI platform. A classic example of a product that was taken to market and then overtaken by the technology progress: Better cloud based solutions, the rise of neural networks and an inability of IBM to truly innovate with Watson (or a successor), make Watson a sad story. It is also a lesson learnt that no matter how much marketing you put into this (IBM coined it cognitive, had a World of Watson conference series). The saddest part for me was the fine line between truth and fiction that IBM executives tangled with closely with – when it came to Watson. But at the core the problem for Watson was again the IBM disease…. (more later).

For some time, IBM was playing well in the PaaS space with BlueMix (see e.g. here), it had its height from 2014-2016 – but then went away as well, as IBM did not manage to make the transition of its Rational install base to the BlueMix platform and it realized that a truly successful PaaS platform would compete with many of it software partners… here is a hint on the IBM disease… again.

More recently IBM has bet more on block-chain, that for some time was the third leg of the IBM stool. But deployments are slow, uptake is not as expected, and things have gotten more quiet around block chain in the last 12 month at IBM.

And then IBM had a large software portfolio of some promising and less promising software assets (yes, IBM sold HCM software with Kenexa, better know as Brassring). How that software portfolio held together, what the synergies were and the master plan – all that went away when legendary IBM software executing Steve Mills retired. Since then the software portfolio has been in disarray, lacks leadership and overall encompassing strategy that at least gives an idea what the 'endgame' will be.

To remain relevant, Rometty had to do something big, and the 'big' thing was the Red Hat acquisition. Enters also the new CEO, Krishna, who was the key architect of the acquisition. The least discussion about the big impact of the acquisition was the price – 34B (read my take here). Red Hat itself was struggling with its new investments to make up for the shrinking RHEL base. The most promising was AppDev – which has gone missing again (likely for the same reasons BlueMix never took off). The next one was the PaaS platform OpenShift, where in October 2017 Red Hat laid the foundation to its multi-cloud support with partnerships with AWS and Google (see here). Curious on the next bet? OpenStack and we know that one went sore for a lot of players. So a lot of expectation is now on creating a hybrid cloud business around the Red Hat OpenShift platform.

IBM's 'disease'

I have attended north of 50 IBM events and conferences in the last 7 years. The company has extraordinary customers, talented people, deep pockets… but it never built truly great software products. IBM felt like an unfinished opera: Act I was hardware, Act II was services and Act III would have been … software. But the products – as solid as they were, as much they were shouldering true enterprise load – they never were truly really great. Close software assets like db2 and WebSphere were more or less thrown out for the next big thing, much to the surprise of customers and other partners. My eye opener was attending the IBM Alliances Insights event August 15th 2016. I always knew that IBM had a ton of consultants… but learning that over 100000 IBM employees made their living with only three partners (SAP, Oracle and Microsoft) made me realize that IBM more than ever, was stuck in the services player category. Simply put – product that were 'good' at IBM would never have seen the light at e.g. Microsoft, Oracle, Salesforce or SAP. They were to service intensive. But for IBM that was ok. Truly great products, that did not need a consultant (e.g. think of BlueMix) were more of a problem. The return of R&D did not translate into a revenue from consulting engagements. And IBM did not understand how to maintain and expand truly great software products (like db2 and WebSphere). The good news is that the hybrid cloud consulting business is immune against the IBM 'disease'. It takes little armies of consultants to get started and remain operating in hybrid cloud… but if enterprises want to operate that – remains a huge and big question. 
Arvind Krishna in an interview with CNBC on April 6th 2020

Krishna has his work cut out

Almost in a position of where the board might have said, you sold us the deal – now make it real – Krishna has to deliver. But to get 34B back from hybrid cloud licenses (small) and professional services (small) is a huge challenge. IBM sees a 1.2T US$ revenue opportunity in hybrid cloud, but it has its set of competitors. AWS, Microsoft and Google have their hybrid offerings – and largely without the need of legions of consultants. IBM may position itself as the hybrid cloud vendor that takes care of the challenging, more service needy clients. In the meantime, the product centric vendors will reduce the need for professional services with every new product release. While having Accenture, Atos, Cap Gemini, Deloitte, KMPG, etc. help them implementing. Mapping progress of the OpenShift platform vs the leading cloud offerings for hybrid cloud will allow to gage IBM's long term success.

Oh and then there is AI. In its initial interview on CNBC, Krishna pointed out that work is being done with Watson (he is still there) Assistant and he was positive on the growth of digital assistants. But no matter how promising digital assistants are – they cannot fuel an IBM as we know it today. So IBM will have to provide a major investment (or acquisition?) in the AI space to be a player. With its fumble on its own cloud infrastructure – it will be a hard market get (back) into. And the challenge was laid in Whitehurst's portfolio with having Bob Lord / Cognitive Applications report directly to him.

Speaking of cloud – IBM's future is literally spilled all over the org chart. Whitehurst and team need to bring OpenShift to the IBM product teams. It makes sense that strategy reports to him. Rob Thomas is also reporting to him with responsibility "security, hybrid cloud, AI and Red Hat synergy" (if you go, Huh? This is word for word from press release). in charge of doing that for the data and cloud products, reporting to Whitehurst. But the platform will be developed under new Red Hat CEO Paul Cormier. Oh yes and then Howard Boville joins from Bank of America (where he was CTO) as SVP for Cloud, overseeing IBM Cloud. Pressure will be also on Bridget van Kralingen who has to deliver the revenue numbers. All of them will need Krishna to call the shots to succeed pretty much on a daily basis. If one of the many strategic IBM customers want and need something, it quickly gets complicated and Krishna will have to play interference.


If Krishna wants to outlive the short tenure of other short tenured IBM CEOs (T. Vincent Leason had 2 years from 1971 to 1973) and John Opel (1981-1985) (see the list of IBM CEOs here) he needs to find a strong second and third revenue stream for IBM. If this still can be AI / Watson is under question. Financially IBM can't make any too big acquisitions, and merging is out of the question (at least for now). Revitalizing the software portfolio, and what is left of it, is an option, especially as Krishna comes from that group. The board gave him a bad setup as giving right away the president position to Whitehurst. Usually IBM CEOs keep that title and it is rarely bestowed (who was the last IBM President that was non CEO – if you know – I will gladly here, it broke Google).On the flip side, if Whitehurst fails, Krishna may be around longer. If Whitehurst succeeds, then Krishna may end up with a tenure with a time between Leason and Opel. Krishna is a smart, humble and likable personality, this will all play in his favor in his new role. Expect to see the by now typical synergies with other key Indian CEOs (Nadella at Microsoft, Pichai at Google and Narayen at Adobe) – which will certainly help.

On the concern side, all is setup for a repeat of the SoftLayer story: Only the 'in house' team has a new head with Boville. But expect the battle between the traditional IBM cloud and Red Hat offerings. Krishna needs to clarify that fast, as customers and partners need to know where IBM is heading. A 10-15 time higher price tag (rumor was that SoftLayer was 2B+ back then) does not guarantee success. Part of it is the customer – why move your cloud offering, if the new offering does not offer enough tangible benefits. A key innovation dilemma that Krishna needs to address, given the more conservative nature of IBM customers and their mission critical workloads.

The opportunity for IBM is multi-cloud. Enterprises do not want lock-in. The question is – how much services do they need for operating in a multi-cloud model and… if the IBM 'disease' will strike again, the sooner Krishna can transform IBM into a "software first" mentality and deliver on it – the better for IBM, its customer, its shareholder and its employees.