Release timing and Office products -- how fast is fast enough?

re: Release timing and Office products -- how fast is fast enough?

PatriotB — Fri, 04 Nov 2005 02:36:06 GMT

It's interesting to see the products & release dates in a table like that. Am I the only one who thinks Office 2003, released in *November* 03, should have been Office 2004? Or VS 2005 should have been VS 2006? (It's kinda strange when 2006 products come out before 2005...)

re: Release timing and Office products -- how fast is fast enough?

steven_sinofsky — Fri, 04 Nov 2005 03:00:21 GMT

Well I don't name the products :-)

Cars do this all the time -- sometimes they want a jump on the next year and sometimes they want to keep the car in teh current year.

Regardless we support the product for 10 years based on the release date not the name.

--Steven

Innovation

Chris Bettin — Fri, 04 Nov 2005 06:55:45 GMT

I think Professor Iansiti could be right about innovation. Especially when the impact is negatively perceived. If you don't have the right deployment, timing, marketing, or audience then your invention is not going to be a smash hit.

I always enjoy reading your blog and hope you keep it up!

re: Release timing and Office products -- how fast is fast enough?

steven_sinofsky — Fri, 04 Nov 2005 08:26:30 GMT

Thanks for the kind words! I will pass along your comments to Prof. Iansiti as well.

re: Release timing and Office products -- how fast is fast enough?

AnOfficeDev — Fri, 04 Nov 2005 10:15:34 GMT

Are you saying we need to have these big monolithic releases since we need to throw lots of features at the user and hope that some of them resonate? Could we change the way we develop, market, and sell the product?

It would be nice if we built more modularized software and could allow for a new "gear shift" to be popped in.

re: Release timing and Office products -- how fast is fast enough?

Fadi — Fri, 04 Nov 2005 11:08:44 GMT

Steven - Very informative post ... I work in MS but know nothing about this...It was worth waiting all this time for this post. Hardik, i hope you're satisfied too.

AnOfficeDev - Couldn't understand what you meant by "would be nice if we built more modularized software and could allow for a new "gear shift" to be popped in" Can you clarify in few more words?

re: Release timing and Office products -- how fast is fast enough?

Fadi — Fri, 04 Nov 2005 11:15:57 GMT

Steven - Why don't you call it Office VISTA and benefit from a 2in1 VISTA campaign. From a marketing standpoint, when you are promoting for "VISTA", you'll be targeting office and windows at the same time thus saving money. Establishing a brand name in the minds of the consumer is very expensive.

Not to forget that having Office VISTA and Windows VISTA drives the customer to get both of the products at the same time since, when having the same release name, they look complimentary.

re: Release timing and Office products -- how fast is fast enough?

steven_sinofsky — Fri, 04 Nov 2005 16:22:56 GMT

Hi there AnOfficeDev -- you know if you are on the team, just step outside your office and head about 100' North and stick your head in my office and ask this question in person. Or maybe today when I drop by the office dev all-hands you can ask this question. I will know it is you and we can make it our secret :-)

I think there are two things:
* Is our software modular?
* If it was modular, why not release features more "one at a time" rather than in big releases

I tried to hit the second, but I did assume the first. I think our software is incredibly modular. If fact the modularity is one thing that contributes to the wide variety of uses and the flexibility that customers enjoy. There are a myriad of ways that you can "plug in" to Office at all sorts of different levels. One of my favorites is that you can switch the user-interface language on the fly, or for that matter the help system can be another language, and you can edit any number of languages (there are lots of features that are required for all the different scripts of the world) -- all without a recompile, relink, or even touching any executable code.

The trick with modularity is that often what seems cool and modular turns out not to be the way you want to extend something in 10 years, which is why during our M0 we are always investing in rearchitecting. One example from the "physical" world is what it is like to set up wiring in a house (or in our case a condo). I spent a huge amount of time on wiring -- I have a great patch panel and home run. I switched DSL providers and it took 10 seconds to patch in a different demarc line and switch over my network. I wanted to change a fax machine to a third line and it took 5 seconds. But I dread the day we might need a 5th phone line (no chance since we only use 2 today but never say never) since my modular system can't handle that. Or even worse what if there is a move to fiber--we have some conduit but not to every room. Software is no different -- you can make it modular to handle a set of circumstances, and for those it works amazingly well. When I was working on VC++/MFC we had a rule which was not to add extensibility unless we used it ourselves, since that way you could prove you were modular for at least some usage. But still customers found things they couldn't accomplish. It is amazing the things that can be easily added to Office and done without a lot of work because we have a great architecture. But like any modular system, it is not "infinite" and time introduces new needs and new scenarios to rearchitect for.

But should we use the modularity we have to do smaller releases? That is not a technical question as much as a sales and marketing question, or more appropriately "would it sell" question, or perhaps a product planning question in terms of up front choices. This one I liken to a movie studio trying to only make hit movies, or an financial manager only picking winning stocks. After all, why waste your time making features that won't "sell". It just isn't so simple--I wish I could think of another way to say that. Some might agree and say that is precisely why you should just release smaller brand new things more frequently and get feedback -- but then that starts to look a lot like "churn" and a lot less like "customer satisfaction". Even our service packs, which make the product way better need to be spaced out because people don't like that much change.

It gets back to "innovation = invention + impact". To have impact something needs to be big and broad. Small improvements are great. A few small improvements might get you a pop. But to really move the needle you need to complete a "scenario" and solve a bigger problem for customers and that takes a lot of invention, some of which won't pan out. Now in some "new" areas or in some hyper markets there is so much focus on invention that it becomes the currency of the day, even without the impact. I think in a "bubble" environment those in the know focus on invention, even though that many people might not be impacted by it. You see that today with many "2.0" things out there--lots of excitement, but exceedingly small numbers of people are impacted by the work. Maybe some of them will be huge, maybe not. It is tough to know. So that is why VCs have dozens of small companies all trying new things. For that same reason, our releases try to address a broad set of customer problems.

Now unlike the "go out of business" option for VCs, it turns out that (as Jensen is going through on his blog) even features that one person does not like in Office are still used by a lot of people. A feature used by 0.5% of our customers in a release, still means that perhaps 500,000 people will be using that feature (based on installed base and historic upgrade rates, not including new users). That is a lot of IMPACT!

So AnOfficeDev, feel free to stop by any time and we can talk more!

--Steven

re: Release timing and Office products -- how fast is fast enough?

Hardik — Fri, 04 Nov 2005 17:04:31 GMT

steven:-sir this is a nice post again.you seems to cover all the long silence in one post ;)
fadi:-yes i really enjoyed reading this blog and jobsblog :) i daily check this two blogs!!!

re: Release timing and Office products -- how fast is fast enough?

Fadi — Fri, 04 Nov 2005 20:18:20 GMT

Steven - Can we visit your office too?... or are you practicing discrimination on your blog? ;-)

re: Release timing and Office products -- how fast is fast enough?

EJ — Fri, 04 Nov 2005 22:16:54 GMT

While this is all a very cogent,detailed, and self-contained account of what I am sure is a huge orchestration issue, it begs the meta-question of "what is the true technical innovation in any of the recent or forthcoming Office versions that justifies 4,000 developers working on it?" At $150K/yr per developer, that is an annual cost of $600M; over 3 years that is nearly $2B of cost. Will we see $2B of technical innovation in Office 12? Or would we be better off spending $2B somewhere else?

re: Release timing and Office products -- how fast is fast enough?

Fadi — Fri, 04 Nov 2005 23:03:31 GMT

EJ - I like your accurate calculations. You just missed adding the coffee and toilet paper consumption cost.

150K/year for each developer! Jack Welch makes a 100K!.

You can't see $2B in innovation! I can't either, cause no one can. Innovation can only be seen through the positive impact produced on people and businesses using office around the world. If you can’t see that, go read some case studies on Microsoft.com.

You may also need to realize how delicate and serious it is to innovate in a product used by 400,000,000 users around the world.

re: Release timing and Office products -- how fast is fast enough?

steven_sinofsky — Fri, 04 Nov 2005 23:22:05 GMT

A perfectly reasonable question from an intellectual point of view, but one that could be asked about any endeavor.

There are only 700 developers on Office though.

I think that the only quantitative measures are patents and revenue. I think by both of these we do a very good job justifying our product development investment. The ROI of our investment in building Office is world-leading I believe.

The positive impact produced by Microsoft Office is pretty incredible. I do wish I could measure it--economists struggle with measuring the economic impact of productivity improvements much more than I ever could. I just stick by what a Wall St. customer once told me -- "we make more money from Excel than Microsoft does". That's innovation and a return on investment.

re: Release timing and Office products -- how fast is fast enough?

jumbo — Sat, 05 Nov 2005 13:21:50 GMT

Please don't make the mistake of VS2005 and ship too early with lots of showstopper bugs. Whoever was in charge of that one really lost the plot.

It's a real pain to use, it crashes all the time and it corrupts data mercilessly.

re: Release timing and Office products -- how fast is fast enough?

steven_sinofsky — Sat, 05 Nov 2005 21:39:39 GMT

Jumbo, I'm sorry you are having that experience. If you have any specifics you would like to report please mail them to me using the contact form.

re: Release timing and Office products -- how fast is fast enough?

Fadi — Mon, 07 Nov 2005 12:19:29 GMT

Steven, do you often interact with Billg during your work?

re: Release timing and Office products -- how fast is fast enough?

Doug Mahugh — Sun, 13 Nov 2005 03:25:07 GMT

When considering Office's development costs, let's not forget all of the developer-oriented code in Office that end users never see. For example, the new XML formats -- huge work to implement, but the average user can't see a thing. And the new APIs in Office 12 for adding 3rd-party functionality to the ribbon, task pane, and so on -- that stuff requires a lot more development work than adding a few bells & whistles to the UI.

As for whether this investment is the best way Microsoft can spend $2 billion (or whatever the number really is), that's a hard question to answer with any precision, like any investment decision. It's clear that people are going to continue using some type of product like Word, Excel, etc, to do their work, whether it's Office or some competitor on the desktop or the net. And it's also clear that we (Microsoft) need to figure out how to remain the number one player in that space.

I'm excited about the developer focus of some of the new features, and how those features will drive ISVs to integrate Office server functionality with their apps to an unprecendented degree. If the investment in Office development can lead to Office being perceived (by developers) as a viable platform for building applications, in addition to its perception as (by end users) as a high-quality productivity suite, then I think we'll look back on it as a wise move.

re: Release timing and Office products -- how fast is fast enough?

Ilana Davidi — Mon, 14 Nov 2005 23:44:24 GMT

Ah, the old "software is easier to change than hardware" argument. We can never tire of that one. =) Your analogy of comparing changing a software product like Office to changing the layout of a magazine rather than the content is a very good one. People who do not work on software can often not understand the concept of infrastructure for something they cannot touch.

Thanks for another informative and detailed article!

PM at Microsoft

Steven Sinofsky's Microsoft TechTalk — Fri, 14 Apr 2006 00:45:29 GMT

While at Stanford this week I was asked by a number of PM (program manager) candidates to talk about...

Ezines and Web Magaizines » Steven Sinofsky’s Microsoft TechTalk : Release timing and Office …

Tue, 29 Apr 2008 09:27:39 GMT

PingBack from http://ezinefrontpageblog.info/steven-sinofskys-microsoft-techtalk-release-timing-and-office/

Steven Sinofsky's Microsoft TechTalk : Release timing and Office products -- how fast is fast enough?

Dating — Sun, 01 Jun 2008 18:08:55 GMT

I've been asked quite a bit by people interviewing at Microsoft about "software release". This post will talk about our strategy around product releases for Office products, since we have been pretty consistent over the years and this represents

Program Manager (PM) at Microsoft

Microsoft's JobsBlog — Wed, 18 Jun 2008 01:02:07 GMT

Republished from Steven Sinofsky's PM at Microsoft , TechTalk, Dec 2005 While at Stanford this week I

Steven Sinofsky s Microsoft TechTalk Release timing and Office | Paid Surveys

Steven Sinofsky s Microsoft TechTalk Release timing and Office | Paid Surveys — Sat, 30 May 2009 02:38:44 GMT

PingBack from http://paidsurveyshub.info/story.php?title=steven-sinofsky-s-microsoft-techtalk-release-timing-and-office

Steven Sinofsky s Microsoft TechTalk Release timing and Office | debt solutions

Steven Sinofsky s Microsoft TechTalk Release timing and Office | debt solutions — Tue, 16 Jun 2009 03:48:11 GMT

PingBack from http://debtsolutionsnow.info/story.php?id=301

Steven Sinofsky s Microsoft TechTalk Release timing and Office | low cost car insurance

Wed, 17 Jun 2009 07:24:39 GMT

PingBack from http://lowcostcarinsurances.info/story.php?id=5196