Having spent several years watching IBM up close and personal while working on OS/2 at Microsoft, I breathed a sigh of relief when IBM bought Lotus. It was clear to me at the time that IBM would spend a ton of money and effort on Notes, but that every worthless "connectivity to OS/360 via CICS" feature that they crammed in would just make Notes less focused, more resource hungry, and less of a long term competitor. It was such an expensive and important acquisition that IBM bureaucracy could not help but smother it. IBM had a *lot* of VPs to feed.
Totally agree that the role of the directory was crucial to Microsoft's success, of course.
I sensed that for sure and read about it enough. So difficult for the team.
What's amazing though is that the directory did not carry over into 365 as an asset in remotely the same way. Look at the now public and private companies in the space that really dominate (Okta/Auth0, Yubi, etc.).
The directory, in the large sense, includes two very distinct services: identity ("who is it that this client claiming to be?") and authentication ("are they actually who they claim they are?"). In AD terms that's LDAP and Kerberos, two architecturally incomplete specifications which we brought together like the chocolate and peanut butter in a Reese's Cup.
In the corporate/organization setting both id and auth are super important, but that's in part because the organizational setting gives everyone an implicit right to know who everyone else is. Displaying a list of identities in the org is perfectly ok. In the public cloud world, though, users have no right to see who else is out there unless they're actually interacting with another person, and even then you have no right to extra identity details about them, just to what they disclose and the fact that the person you're dealing with today is provably the same person you were dealing with yesterday. Therefore in the cloud identity (directory narrowly defined) is relatively less important and authentication is more important than in an org setting. Okta, Yubi, etc, are all authentication vendors, not identity vendors.
Technically of course you are correct. But in a broader sense, Okta customers view the product as identity. When a human joins a company, the first thing they get is an Okta identity (and Okta drives on-boarding).
This is not unlike the battle Microsoft had with SAP over which was "on top". We used to talk about this as what comes first, an SAP identity or AD credentials. For a long time SAP came first because that was where you got paid. Over time what happened was that to get email you needed AD and so AD came first. You can see that in the direction of sync, and how it used to be from AD to SAP and then went to SAP to AD as the main source.
Today when customers use Okta or competitors that's where they start.
Well, I've had a to-do to subscribe to your writing and come read this for months, and I'm genuinely surprised nobody else from the Notes side has done so in the four months since you published.
I do find it funny you write this
" In other words, all this stress and talk about how great Notes is and how behind Microsoft is and yet Microsoft won. Crazy."
Microsoft won in email. Well, mostly, Google certainly has taken some of that market share by now as well. But Outlook did eventually displace Notes in corporate messaging.
But here in 2022, thirty two years after Notes was first introduced, it is still (as a HCL product) generating hundreds of millions of dollars in revenue and is still used by over 10,000 companies around the world.
Microsoft eventually did with Notes exactly what you describe in this article. Defined it as email, a single category, bundled the email product with must-buy Windows CALs and Office, and argued it was therefore a no-brainer to use the "free" Microsoft product. There was no way to beat the Notes custom applications, despite numerous attempts to do so, so eventually MS - and Salesforce, and every other "Notes killer" - stopped trying. And those applications keep chugging along.
There is no doubt, some of the wounds on Notes were highly self-inflicted. I ran the business through many of them, directives from way up the food chain. A classic example that I'll share was when we were told we could not invest in any sort of predictive text approach in cloud email (for a new front-end called Verse, but it was still Notes) unless it was based on Watson, and the Watson team wanted a per-sentence internal transaction fee that would have completely eroded the profit margin of the **entire Notes business**.
I would love to some day be part of the "someday someone will interview all the principles and write a full 360-degree account". I've threatened to write a history book on the whole of Lotus Notes taken from the IBM/Lotus point of view; I did some Lotusphere presentations about it but there are so many stories that I can now share (good and bad) with the distance of history. But the product is still alive, so it has never quite felt like the right time. Maybe, someday.
Hello, I did not mean for the bravado to be a substitute for sales numbers. As we know in technology, there's when it comes to winning there's leadership position, market share, revenue share, and ecosystem health to name a few. When it comes to marketing or managing, we pick the ones favorable to us in some context. Notes has been an enduring product. The leadership position it had in many dimensions did not endure as long relative to Exchange/Outlook.
PS: I know many of the original team saw this post. Sorry the comments were not posted here. We're all learning substack and what works best.
I'm a serious Alan Cooper and Charles Petzold fan. I would say that the Chandler effort is somehow a failure in this "but x is not Notes" difficulty, at least as depicted in "Dreaming in Code." Novell wanting GroupWise to provide such an equivalent was saddled with WordPerfect and Quattro Pro flawed integrations and distractions with model changes (e.g., Bento) and confusion of directory services with databases. This reminds me of watching the early email list of the Euclid project and its preoccupation with plug-ins and component chaining. Chandler changed its tooling and internal architecture at least twice, and there seemed to be no way to ground the project around user functions rather than open-source tantalizations.
In this context, I would say that Microsoft managed to pull away from the cliffs on the Outlook, Office, Exchange, Sharepoint, Azure winding trajectory. Having big and demanding enterprise customers may have been a saving grace, along with having integrated Office well. (That FrontPage and the Site Server model was sacrificed to whatever infighting imagined Sharepoint to be the world's CMS solution is something that has me be wary of fledgling off-shoots in the Microsoft quiver.)
I had the original public beta of Outlook and it was a winner. I also had Ecco and was not dismayed when the producer abandoned the product, wisely concluding that Outlook was going to eat their market. I could never figure out the conceptual model for Ecco, and you needed to have one to make it work well and not do disastrous deletions of cruft. With Outlook, my use cases worked fine and I didn't have to plumb the mysteries of the PST and the forms mechanism. Truth be told, it is the Outlook client that keeps this household on Microsoft 365.
Well, synchronizing a PST over dial-up between my home system and work desktop was fairly frightening. I can't remember how I managed to move it out of my work computer, since PARC on the Internet but corporate was on XNS for a short while longer :). These days, use of live/hotmail/outlook.com to synchronize calendars and contact lists with my spouse and also with our smartphone devices is wonderful. I do miss my Luna though.
Very nice article. I have been told by an MS executive that in addition the Enron downfall had the happy unexpected effect to greatly accelerate the adoption of Exchange in corporations, supposedly because the server management features were closer to SOX requirements, but I was not able to correlate this story with reading elements or pinpoint which features exactly. From what I know Lotus had centralized email management before 2000.
Or maybe that person was referring to customers with light email solutions that Microsoft would had lost for a long time if Lotus was the only option at the time the SOX was finalized.
Compliance became a significant source of strength for Exchange. Security as well. The addition of S/MIME was the only way to win the DoD contract, for example.
About 6 months after I joined MSFT – a year after the launch of the iPhone - I knew enough about the people there to ask this question: “These are some of the smartest, most organized people I have ever worked with. So why is it that Apple develops the iPhone, and we develop the Windows Phone 6?”. This was right after the SteveB partner meeting where he put up a competition slide showing Nokia as #1, RIM as #2, MSFT as #3, and Apple as #4 but fast-growing (Android wasn’t a factor then), and then somehow concluded that despite being surrounded in hist list by companies building fully integrated systems instead of just licensing the OS like we did, that our model was best. Of course, I understood confirmation bias at a company that had generated enormous wealth using that model, but it was very unconvincing.
I puzzled over it later, and then determined that we were thinking of this product as a list of features rather than what the customer saw when they used one. As you said: “the underlying technologies of a product might not always be the most important element.” When I read that I scribbled “This explains the phone” in the margins of my printout of this post.
Perhaps it was my own confirmation bias but after concluding this I started to see this phenomenon across some of the company’s teams and products. I often wondered if other execs saw what is saw. I talked about it with some other execs and most of them treated my question as a form of heresy from a new hire rather than consider my premises. Perhaps I didn’t set up my observation well, I thought.
So, I was amazed to read this later in your post: “This pattern is one we saw time and again, phones were probably a good example of that type of failure.”
It’s well-hidden in this post, but to me, this post is the best possible explanation for Microsoft’s failure with smart phones I have ever read.
Such a great post. It should be read by a broader audience – the industry was and still is very puzzled by this misstep. Your post explains it, albeit obliquely in an excellent post about Lotus Notes and MSFT strategy. It should become a B-school business case paper.
Having spent several years watching IBM up close and personal while working on OS/2 at Microsoft, I breathed a sigh of relief when IBM bought Lotus. It was clear to me at the time that IBM would spend a ton of money and effort on Notes, but that every worthless "connectivity to OS/360 via CICS" feature that they crammed in would just make Notes less focused, more resource hungry, and less of a long term competitor. It was such an expensive and important acquisition that IBM bureaucracy could not help but smother it. IBM had a *lot* of VPs to feed.
Totally agree that the role of the directory was crucial to Microsoft's success, of course.
I sensed that for sure and read about it enough. So difficult for the team.
What's amazing though is that the directory did not carry over into 365 as an asset in remotely the same way. Look at the now public and private companies in the space that really dominate (Okta/Auth0, Yubi, etc.).
The directory, in the large sense, includes two very distinct services: identity ("who is it that this client claiming to be?") and authentication ("are they actually who they claim they are?"). In AD terms that's LDAP and Kerberos, two architecturally incomplete specifications which we brought together like the chocolate and peanut butter in a Reese's Cup.
In the corporate/organization setting both id and auth are super important, but that's in part because the organizational setting gives everyone an implicit right to know who everyone else is. Displaying a list of identities in the org is perfectly ok. In the public cloud world, though, users have no right to see who else is out there unless they're actually interacting with another person, and even then you have no right to extra identity details about them, just to what they disclose and the fact that the person you're dealing with today is provably the same person you were dealing with yesterday. Therefore in the cloud identity (directory narrowly defined) is relatively less important and authentication is more important than in an org setting. Okta, Yubi, etc, are all authentication vendors, not identity vendors.
Technically of course you are correct. But in a broader sense, Okta customers view the product as identity. When a human joins a company, the first thing they get is an Okta identity (and Okta drives on-boarding).
This is not unlike the battle Microsoft had with SAP over which was "on top". We used to talk about this as what comes first, an SAP identity or AD credentials. For a long time SAP came first because that was where you got paid. Over time what happened was that to get email you needed AD and so AD came first. You can see that in the direction of sync, and how it used to be from AD to SAP and then went to SAP to AD as the main source.
Today when customers use Okta or competitors that's where they start.
Well, I've had a to-do to subscribe to your writing and come read this for months, and I'm genuinely surprised nobody else from the Notes side has done so in the four months since you published.
I do find it funny you write this
" In other words, all this stress and talk about how great Notes is and how behind Microsoft is and yet Microsoft won. Crazy."
Microsoft won in email. Well, mostly, Google certainly has taken some of that market share by now as well. But Outlook did eventually displace Notes in corporate messaging.
But here in 2022, thirty two years after Notes was first introduced, it is still (as a HCL product) generating hundreds of millions of dollars in revenue and is still used by over 10,000 companies around the world.
Microsoft eventually did with Notes exactly what you describe in this article. Defined it as email, a single category, bundled the email product with must-buy Windows CALs and Office, and argued it was therefore a no-brainer to use the "free" Microsoft product. There was no way to beat the Notes custom applications, despite numerous attempts to do so, so eventually MS - and Salesforce, and every other "Notes killer" - stopped trying. And those applications keep chugging along.
There is no doubt, some of the wounds on Notes were highly self-inflicted. I ran the business through many of them, directives from way up the food chain. A classic example that I'll share was when we were told we could not invest in any sort of predictive text approach in cloud email (for a new front-end called Verse, but it was still Notes) unless it was based on Watson, and the Watson team wanted a per-sentence internal transaction fee that would have completely eroded the profit margin of the **entire Notes business**.
I would love to some day be part of the "someday someone will interview all the principles and write a full 360-degree account". I've threatened to write a history book on the whole of Lotus Notes taken from the IBM/Lotus point of view; I did some Lotusphere presentations about it but there are so many stories that I can now share (good and bad) with the distance of history. But the product is still alive, so it has never quite felt like the right time. Maybe, someday.
Hello, I did not mean for the bravado to be a substitute for sales numbers. As we know in technology, there's when it comes to winning there's leadership position, market share, revenue share, and ecosystem health to name a few. When it comes to marketing or managing, we pick the ones favorable to us in some context. Notes has been an enduring product. The leadership position it had in many dimensions did not endure as long relative to Exchange/Outlook.
PS: I know many of the original team saw this post. Sorry the comments were not posted here. We're all learning substack and what works best.
I'm a serious Alan Cooper and Charles Petzold fan. I would say that the Chandler effort is somehow a failure in this "but x is not Notes" difficulty, at least as depicted in "Dreaming in Code." Novell wanting GroupWise to provide such an equivalent was saddled with WordPerfect and Quattro Pro flawed integrations and distractions with model changes (e.g., Bento) and confusion of directory services with databases. This reminds me of watching the early email list of the Euclid project and its preoccupation with plug-ins and component chaining. Chandler changed its tooling and internal architecture at least twice, and there seemed to be no way to ground the project around user functions rather than open-source tantalizations.
In this context, I would say that Microsoft managed to pull away from the cliffs on the Outlook, Office, Exchange, Sharepoint, Azure winding trajectory. Having big and demanding enterprise customers may have been a saving grace, along with having integrated Office well. (That FrontPage and the Site Server model was sacrificed to whatever infighting imagined Sharepoint to be the world's CMS solution is something that has me be wary of fledgling off-shoots in the Microsoft quiver.)
I had the original public beta of Outlook and it was a winner. I also had Ecco and was not dismayed when the producer abandoned the product, wisely concluding that Outlook was going to eat their market. I could never figure out the conceptual model for Ecco, and you needed to have one to make it work well and not do disastrous deletions of cruft. With Outlook, my use cases worked fine and I didn't have to plumb the mysteries of the PST and the forms mechanism. Truth be told, it is the Outlook client that keeps this household on Microsoft 365.
Thank you! You might her been in a small group of people that thought outlook was a winner with the beta/97 :-)
I love Dreaming in Code! https://smile.amazon.com/Dreaming-Code-Programmers-Transcendent-Software/dp/1400082471/
Well, synchronizing a PST over dial-up between my home system and work desktop was fairly frightening. I can't remember how I managed to move it out of my work computer, since PARC on the Internet but corporate was on XNS for a short while longer :). These days, use of live/hotmail/outlook.com to synchronize calendars and contact lists with my spouse and also with our smartphone devices is wonderful. I do miss my Luna though.
Very nice article. I have been told by an MS executive that in addition the Enron downfall had the happy unexpected effect to greatly accelerate the adoption of Exchange in corporations, supposedly because the server management features were closer to SOX requirements, but I was not able to correlate this story with reading elements or pinpoint which features exactly. From what I know Lotus had centralized email management before 2000.
Or maybe that person was referring to customers with light email solutions that Microsoft would had lost for a long time if Lotus was the only option at the time the SOX was finalized.
Compliance became a significant source of strength for Exchange. Security as well. The addition of S/MIME was the only way to win the DoD contract, for example.
This post was just excellent, Steven.
About 6 months after I joined MSFT – a year after the launch of the iPhone - I knew enough about the people there to ask this question: “These are some of the smartest, most organized people I have ever worked with. So why is it that Apple develops the iPhone, and we develop the Windows Phone 6?”. This was right after the SteveB partner meeting where he put up a competition slide showing Nokia as #1, RIM as #2, MSFT as #3, and Apple as #4 but fast-growing (Android wasn’t a factor then), and then somehow concluded that despite being surrounded in hist list by companies building fully integrated systems instead of just licensing the OS like we did, that our model was best. Of course, I understood confirmation bias at a company that had generated enormous wealth using that model, but it was very unconvincing.
I puzzled over it later, and then determined that we were thinking of this product as a list of features rather than what the customer saw when they used one. As you said: “the underlying technologies of a product might not always be the most important element.” When I read that I scribbled “This explains the phone” in the margins of my printout of this post.
Perhaps it was my own confirmation bias but after concluding this I started to see this phenomenon across some of the company’s teams and products. I often wondered if other execs saw what is saw. I talked about it with some other execs and most of them treated my question as a form of heresy from a new hire rather than consider my premises. Perhaps I didn’t set up my observation well, I thought.
So, I was amazed to read this later in your post: “This pattern is one we saw time and again, phones were probably a good example of that type of failure.”
It’s well-hidden in this post, but to me, this post is the best possible explanation for Microsoft’s failure with smart phones I have ever read.
Such a great post. It should be read by a broader audience – the industry was and still is very puzzled by this misstep. Your post explains it, albeit obliquely in an excellent post about Lotus Notes and MSFT strategy. It should become a B-school business case paper.
This post and a comment from me might be helpful wrt phones.
https://christianhern.medium.com/the-war-microsoft-should-have-won-65d836aa2358
https://medium.com/@stevesi/it-is-worth-recognizing-that-microsoft-tried-to-recreate-the-pc-ecosystem-with-mobile-e114a1958c18