The Perils of Platform As A Service

About This Blog

This site is for  entrepreneurs.  A full RSS feed to the articles is available.  Please subscribe so we know you're out there.  If you need more convincing, learn more about the site.



And, you can find me on Google+

Connect on Twitter

Get Articles By Email

Your email:


Blog Navigator

Navigate By : 
[Article Index]

Questions about startups?

If you have questions about startups, you can find me and a bunch of other startup fanatics on the free Q&A website:

Subscribe to Updates


30,000+ subscribers can't all be wrong.  Subscribe to the RSS feed.

Follow me on LinkedIn


Current Articles | RSS Feed RSS Feed The Perils of Platform As A Service


First off, let me say I'm very impressed with I admire the company considerably. My startup, HubSpot, is a customer. We're also about to become a partner so we can integrate our marketing product with the SFDC CRM/sales application and create some benefits for our customers.

As a CRM applications company, I think is simply brilliant and the company deserves much of the success it has achieved. I even like the fact that they built an API so that third-parties can create value around the core CRM application. This is just smart as it further extends their reach and gives the ability for customers and partners to create more value.

However, I'm a bit troubled by the recent announcement of, SFDC's new "platform as a service" offering. My concern is not with the actual underlying technology, as I don't know much about it. I've never written a line of Apex code in my life (this is SFDC's proprietary language introduced as part of the hosted apps on AppExchange). My concerns are primarily strategic in nature and centered around software entrepreneurs looking to build businesses on top of My guess is that startups are the initial primary audience for as I don't believe large, established enterprises will be buliding or porting apps to anytime soon.

So, if you're a software entrepreneur, here are my thoughts on the tradeoffs of building your shiny new startup on top of

1. Lack of an ecosystem: I think the ecosystem surrounding a platform is as important as its underlying technical merits. Where will you find other developers? Books? Training? Tools? Components? If you have to initially rely on SFDC and it's small pool of early partners, chances are, you're going to pay higher than average costs for all of these things.

2. Upper Limit On Growth: Lets say your idea really is as brilliant as you think it is. How far do you think you can take it on if you're running on $10M? $100M? The next Facebook? I'm going to argue that there's an intrinsic limit on how big your company can get while running on I clearly have no data to back up this argument (yet), but see point #3 below for some rationale.

3. Royaties: So, what will being on cost you? Unlike successful platforms of yesteryear (like DOS, Windows, Mac or the Internet), has a royalty. You're going to pay $25/month/user to run your app on it. I like the simplicity of this, but it does limit the kinds of apps you can build. There are lots of software businesses out there where $25/month/user would likely be over half the revenue generated. Sure, you get infrastructure, distribution help (in theory) and a shorter development cycle. But, this comes at a price.

4. Price Controls (or lack thereof): I would worry that in the long term, would have significant "power" over the price -- and would likely find ways to exert that power over startups. I'm not saying that they're necessarily going to *raise* prices, but they're likely going to take a larger fraction of the "market value" for the capabilities they're bringing to the table. So, 5 years from now, when processors, storage, bandwidth and other infrastructure components are even cheaper, there's no requirement that SFDC has to reduce its prices to match "market". Once you're on, you're on. Another alternative would be to come up with "Enterprise" (basically a segmentation and price discrimination strategy). The features included in the "Enterprise" platform would be just those those successful on the platform really have to have.

5. Focused Competitive Power: Since this is "platform as a service", SFDC would have considerable power over the startups that run on it's platform. Back in the day of DOS/Windows, there was suspicion that Microsoft went out of their way to ensure that apps like Lotus 1-2-3 and WordPerfect had "trouble" running on their latest OS. But, back then, the product update cycle was pretty long and there were literally thousands of developers using the platform that Microsoft didn't necessarily know about. Even if they Microsoft hated your guts (because you were competing with them), it would take them some time and effort to really single you out and do damage. They couldn't easily change the platform to make your life (individually) hard. But, with, the power to change is immediate. If you failed to negotiate a new partner agreement, pay your royalties, or (gasp!) wind up in a situation where SFDC really liked your market, they could (technically) turn you -- and your customers -- off in a second. There's nowhere to hide. I'm not saying they'd do this, but simply having the *ability* to do this gives them a fair amount of power and future negotiating leverage.

My point is, though I find the "platform as a service" idea interesting, I'd be a little leery of actually putting all of my startup eggs in that basket. I'm sure there will be entrepreneurs, VCs and other folks that will support SFDC in this effort and work to make it successful. But, for now, my bet is that SFDC won't be nearly as successful in the broad "platform as a service" market as they were in the CRM apps business. It's just a very different game.

This is why I like the Facebook platform. You get the advantage of a large and growing number of users on the platform -- without all the muckiness of having to use their language, tools and technologies. In fact, you could simultaneously build an app that lives on Facebook -- but also runs on it's own, thereby giving you the best of both worlds. Sure, you still have to worry about infrastructure -- but that stuff is geting easier and cheaper.

What do you think? Are you considering spending the next couple of years building something on Am I being overly paranoid here and missing what will prove to be the greatest opportunity since Windows for software developers? Would love to read your thoughts in the comments.

Posted by Dharmesh Shah on Fri, Oct 05, 2007


Valid concerns. ning has a much better approach worth some analysis. You're allowed to take you code and go home. With SalesForce, it is typical lock in, once invested, all you have is hope that salesforce will be smart enough not to poison the well by screwing its partners ..

posted on Friday, October 05, 2007 at 12:10 PM by Berkay Mollamustafaoglu

++1 for Ning - Walled gardens are fine - if the walls are there to keep "the wrong people" [spammers] out. I think when they are used to keep customers captive they plant the seeds of defection... Good post Dharmesh!

posted on Friday, October 05, 2007 at 1:24 PM by Bob Walsh

The link to HubSpot is broken. It contains an extra "http://" at the end. Nice article.

posted on Friday, October 05, 2007 at 3:50 PM by JR

Good post. As SAAS start up developping an application for a small niche, we are currently evaluating where to finally develop and host our software. There are other Saas platforms besides, as Etelos and Apprenda with similar functionality . These two platforms have also their good and bad points, similar to the ones you mentioned. But funny enough, few blogs and magazines talk about them.

Deciding whick platform to use if any, is a difficult decision to make and I would really appreciate if someone has made a comparison among Saas platforms that could help us taking the decision.


posted on Sunday, October 07, 2007 at 2:29 PM by Ramon

Hi All,
Sooner i am coming up with SaaS....which will surly create a good marketplace....currently i am developing prototype...and dont have a proper bussiness model...and also looking out for investors....
--Good article....


posted on Monday, October 08, 2007 at 1:14 AM by Uday

I've always tried to separate value (is this something the customer must have at any cost) from value proposition (value/cost). It's very easy to get into the pricing and costs questions prematurely, and I've worked with many early stage companies that blow through their core worth, assuming that what they've specified or built must have a strong pricing and margin story. My take is that if you can find an application that is valuable, the platform play is a reasonable strategy to get to the market fast.

posted on Monday, October 08, 2007 at 7:21 AM by Ed Wagner

Great post!

Choose your platform vendor very carefully, but I do believe there are good answers to a lot of these questions that don't invalidate the idea of Platforms as a Service. I'm with you that Force raises more questions than it answers.

Here are my thoughts on the problem:

posted on Monday, October 08, 2007 at 6:38 PM by Bob Warfield

For software entrepreneurs these are certainly valid concerns. However, there is an arguably larger market for PaaS consisting of non-programmers -- i.e. web-savvy business users with either specific business requirements or perhaps ideas for entirely new business apps -- who do not have the skills or financial resources to build, manage or even host their own software.

PaaS vendors such as Rollbase - disclaimer: self-promotion) allow non-programmers to quickly build enterprise apps without writing code, for either internal use or distribution to a target market. This kind of DIY approach, rather than being a threat let alone a peril to software developers, opens up entirely new opportunities for businesses and individuals without the resources or know-how to deal with software development. In this context building out a business on PaaS can be an attractive option where there simply was none, or very few, before.

posted on Friday, October 12, 2007 at 7:26 PM by Matt Robinson

I think the days are here now where you have to figure support for multiple platforms into your business model. I don't know what your business is but SF might be the way to go for you when you consider the alternatives. PS - SF might argue #2 with you; they can trot out very large customers as refernces when they need to.

posted on Monday, October 15, 2007 at 10:44 AM by Dennis McDonald

You would be right, if SFDC were not evolving. For now, companies will remain tightly tied to SFDC customers and needs, and Appexchange related business issues. That means you won't built - at the present - a new Facebook or a new Flickr because those are not related to mainstream SFA/CRM needs. You won't be concerned about items 1-5 yet, as long as your company lives within the cocoon, sponsored by SFDC and in fact adding to the original SFDC power. As is evolving, the ecosystem will grow, tieing SFDC into companies larger than it, more powerful, in areas where SFDC has no control, and in an entangling way much different than the usual client/supplier relationship. If at some point, SFDC is tied with a large law firm and a financial company, I don't expect them to have the strength to "play" with touching at customer's data and features. Without going that far, I dont' see why technical differences between past platforms (Microsoft) and Web platforms (SFDC) would be reasons to imagine new ways of circumventing contracts which, I assume, are based on the same legal framework. As to pricing, it has been evolving, I haven't read a reason why it would stop, or continue to be as flexible as it is. Now, the main risk I see comes from SFDC employees themselves, being able to access every customers' data. I'd be scared to be a customer in a highly competitive market.

posted on Tuesday, October 16, 2007 at 1:41 PM by Pierre

Interesting article Dharmesh! Something we could use while we're mapping out our SaaS strategies.

posted on Saturday, December 08, 2007 at 7:26 PM by Alain

Hi Dharmesh,
Good article on PaaS.
I think it will be difficult for PaaS vendors to sell the concept to B2B ISVs. ISVs have often had real problems convincing themselves that is OK to use 3rd party components in their products. These doubts were amplified when the first development tools appeared with proprietary runtime systems.
PaaS moves dependency to a new level. The ISV will be totally reliant on the PaaS vendor for both the tools and the deployment. They are really putting all of their eggs in the one basket. There are positive aspects of PaaS, but for commercial ISVs I think the risk of lock-in to a specific PaaS will make them stop and think long and hard.
On the other hand, there can be an interesting market for a standards-based development tool combined with a PaaS offer. This is the market the Ruby on Rails hosting companies are addressing. When using a standard tool an ISV has the option (at least in theory) of moving from one PaaS vendor to another. They may never use this option, but it will certainly help during the sales process to know that there is a way out.

posted on Wednesday, February 27, 2008 at 1:36 PM by Andrew Biss

Great post Dharmesh. 
My product is SaaS and I have been examining this space carefully. The Amazon cloud allows you to upload a virtual image of your linux products up to the cloud and pay based on how much cpu that you use. And you can flip a switch and be hosted on another continent. I like that. 
With you now have a business partner. That could be problematic. But the risks could be offset by getting to market sooner. If you are really successful, you should immediately implement a backup strategy. As long as you have the domain name you can point to whereever you want. 
One of my questions is how are they going to really make money off me if my product is not about 'users'? 
If I decide to build a real estate management application and I have 3 users and thousands of people searching listings all day long, am I going to be punished with an excess hit switch? 
Frankly, if decided to charge a percentage of what you charge your own client, I think they might hit it off better.

posted on Friday, February 27, 2009 at 3:26 PM by Paul Murray

I'll disagree with you that is mainly suited for startups. Large corporates have joined the band wagon because of the flexiblity and effeciency the platform offers. However certainly needs to work on its pricing models and streamlining of the platform. At first glance it looks more like a maze, all in all it has great potential.

posted on Sunday, March 22, 2009 at 7:09 AM by Ubitmo

Comments have been closed for this article.