<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>
<channel>
	<title>Comments on: Running a SaaS Business Costs Money?!?</title>
	<atom:link href="http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/</link>
	<description>Understanding the Software as a Service Revolution</description>
	<pubDate>Tue, 16 Mar 2010 18:57:17 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5</generator>
		<item>
		<title>By: Eric Novikoff</title>
		<link>http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-45058</link>
		<dc:creator>Eric Novikoff</dc:creator>
		<pubDate>Mon, 12 May 2008 17:43:23 +0000</pubDate>
		<guid isPermaLink="false">http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-45058</guid>
		<description>As an operations services provider for startups (among others) I see this issue cropping up again and again.  Most SaaS startups expect to pay about 10% of revenue for their operations, which as you point out is unrealistically low.  I actually spend a lot of time coaching them on their P&#38;L modeling!  I think the key to remember is that as the business grows, the operations cost per customer can be driven down by tuning the application and leveraging economies of scale in both hosting cost and labor.  The first few customers are going to be the most "painful" from a cost of operations standpoint, and shouldn't be taken as a predictor of the future - unless you fail to attend to optimizing your application deployment over time and of course to growing your business.  Even the best operations cost situation doesn't help if you don't have many customers!  It definitely pays to utilize pay-as-you-go (utility-billed) services such as the ones we offer to reduce the unused resources you have to pay for. This is especially true at the beginning when you might need three full servers to meet reliability and scalability requirements and 3-5 full time people to get the breadth of expertise and round-the-clock response time to run an enterprise-class data center, even though you're only utilizing at most a tiny fraction of all those resources.  You can think of it as a "needs curve" of costs for services you actually need per user which if you fulfilled it by leasing servers and hiring people would have a "costs curve" that went up in big steps, with the area between the costs and needs being unused resources that you're paying for.

Eric Novikoff
http://www.ComputingUtility.com</description>
		<content:encoded><![CDATA[<p>As an operations services provider for startups (among others) I see this issue cropping up again and again.  Most SaaS startups expect to pay about 10% of revenue for their operations, which as you point out is unrealistically low.  I actually spend a lot of time coaching them on their P&amp;L modeling!  I think the key to remember is that as the business grows, the operations cost per customer can be driven down by tuning the application and leveraging economies of scale in both hosting cost and labor.  The first few customers are going to be the most &#8220;painful&#8221; from a cost of operations standpoint, and shouldn&#8217;t be taken as a predictor of the future - unless you fail to attend to optimizing your application deployment over time and of course to growing your business.  Even the best operations cost situation doesn&#8217;t help if you don&#8217;t have many customers!  It definitely pays to utilize pay-as-you-go (utility-billed) services such as the ones we offer to reduce the unused resources you have to pay for. This is especially true at the beginning when you might need three full servers to meet reliability and scalability requirements and 3-5 full time people to get the breadth of expertise and round-the-clock response time to run an enterprise-class data center, even though you&#8217;re only utilizing at most a tiny fraction of all those resources.  You can think of it as a &#8220;needs curve&#8221; of costs for services you actually need per user which if you fulfilled it by leasing servers and hiring people would have a &#8220;costs curve&#8221; that went up in big steps, with the area between the costs and needs being unused resources that you&#8217;re paying for.</p>
<p>Eric Novikoff<br />
<a href="http://www.ComputingUtility.com" rel="nofollow">http://www.ComputingUtility.com</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Matt Robinson</title>
		<link>http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-17389</link>
		<dc:creator>Matt Robinson</dc:creator>
		<pubDate>Sun, 14 Oct 2007 20:26:49 +0000</pubDate>
		<guid isPermaLink="false">http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-17389</guid>
		<description>This is particularly helpful as I embark on a P&#38;L modeling excercise. Have folks here found that CoR is the most useful metric to pay attention to? My current models suggest overwhelming economies of scale and I imagine this should be more or less true for any pure multitenant SaaS vendor. Yet you say "rudimentary evidence for economies of scale" -- curious if there is any evidence to the contrary?

Matt Robinson
www.rollbase.com
Roll your own business apps</description>
		<content:encoded><![CDATA[<p>This is particularly helpful as I embark on a P&amp;L modeling excercise. Have folks here found that CoR is the most useful metric to pay attention to? My current models suggest overwhelming economies of scale and I imagine this should be more or less true for any pure multitenant SaaS vendor. Yet you say &#8220;rudimentary evidence for economies of scale&#8221; &#8212; curious if there is any evidence to the contrary?</p>
<p>Matt Robinson<br />
<a href="http://www.rollbase.com" rel="nofollow">http://www.rollbase.com</a><br />
Roll your own business apps</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: En Avant &#187; Blog Archive &#187; SaaS business blog</title>
		<link>http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-5787</link>
		<dc:creator>En Avant &#187; Blog Archive &#187; SaaS business blog</dc:creator>
		<pubDate>Thu, 24 May 2007 15:38:42 +0000</pubDate>
		<guid isPermaLink="false">http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-5787</guid>
		<description>[...] Running a SaaS business costs money?  Trackback uri [...]</description>
		<content:encoded><![CDATA[<p>[&#8230;] Running a SaaS business costs money?  Trackback uri [&#8230;]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: SaaS Blogs - &#187; On User Density</title>
		<link>http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-4493</link>
		<dc:creator>SaaS Blogs - &#187; On User Density</dc:creator>
		<pubDate>Tue, 08 May 2007 13:54:41 +0000</pubDate>
		<guid isPermaLink="false">http://www.saasblogs.com/2007/05/06/running-a-saas-business-costs-money/#comment-4493</guid>
		<description>[...] My last post discussed cost of revenue related to delivering a SaaS-offering to a subscriber base. Shortly after publishing it, I noticed that Gianpaolo Carraro posted this article, which toward the end discussed the relationship between cost and electricity/space within a datacenter. A while back, Gianpaolo described another metric: tenants per database and he described it as &#8220;density.&#8221; I think that really summarizes the electricity/real estate metric and is the best way to look at any SaaS metric. Drive cost of revenue (and even operational costs) down: maximize your density metrics. As a SaaS provider, you need ot identify any &#8220;Users per X&#8221; or &#8220;Tenants per Y&#8221; and figure out ways to make that ratio bigger. Doing so will improve cost predictability, buffer against sudden cost change (distribution property), and most importantly it will build your economies. Changing these ratios isn&#8217;t easy, however. Most of the time, good density is achieved through advanced application architectures, good infrastructure configuration, etc.  Tags: &#160;   These icons link to social bookmarking sites where readers can share and discover new web pages. [...]</description>
		<content:encoded><![CDATA[<p>[&#8230;] My last post discussed cost of revenue related to delivering a SaaS-offering to a subscriber base. Shortly after publishing it, I noticed that Gianpaolo Carraro posted this article, which toward the end discussed the relationship between cost and electricity/space within a datacenter. A while back, Gianpaolo described another metric: tenants per database and he described it as &#8220;density.&#8221; I think that really summarizes the electricity/real estate metric and is the best way to look at any SaaS metric. Drive cost of revenue (and even operational costs) down: maximize your density metrics. As a SaaS provider, you need ot identify any &#8220;Users per X&#8221; or &#8220;Tenants per Y&#8221; and figure out ways to make that ratio bigger. Doing so will improve cost predictability, buffer against sudden cost change (distribution property), and most importantly it will build your economies. Changing these ratios isn&#8217;t easy, however. Most of the time, good density is achieved through advanced application architectures, good infrastructure configuration, etc.  Tags: &nbsp;   These icons link to social bookmarking sites where readers can share and discover new web pages. [&#8230;]</p>
]]></content:encoded>
	</item>
</channel>
</rss>
