ckluis

There are many who have claimed the marginal cost of software is zero. I essentially agree, but with a slight caveat – the marginal cost of software approaches zero (although it can never truly hit zero).

For a more elegantly written version see: the marginal cost of software

This statement is contentious to some. In fact, I’ve overheard: “sure it works for apps, but what about SaaS software? What about support, maintenance, & infrastructure?”

Let’s start with some loose definitions:

• marginal cost = the cost to produce the next unit
• fixed costs = costs that are not reducible in the short term (lease)
• variable costs = costs that scale with production

Marginal cost includes both fixed and variable costs

In software the cost of software production is primarily fixed – they do not change regardless of how much you sell. If it cost 10 million to build Microsoft Office – there is not much additional variable cost for the 1,000th unit sold. Once you pass break-even you have a growing amount (both in dollars and in percent of revenue) of profit with each unit sold.

In the production of jeans – you have fixed costs (the plant, equipment, and some labor) and variable costs (the materials, some labor, storage, shipping, sales). There will be a sizable chunk of the value of the jeans that is directly related to the production of that particular pair of jeans – also known as the COGS (cost of goods sold).

So let’s dive into the contentious areas – support, maintenance, & infrastructure.

The marginal cost of providing support approaches zero.

The concept of the marginal cost approaching zero – literally means that it becomes cheaper with every additional customer to provide the service. Let’s evaluate that in the context of a small software company in the enterprise space with a handful of large customers:

• the first customer requires a minimum of 5 support staff to cover available hours of operation and the call schedule
• the second customer requires adding a person 6 support staff; this is clearly an increased cost to supporting 2 customers
• the third customer doesn’t require adding any

In this scenario we have a perfect example of the reduced nature of support requirements at a small scale; with each successive customer we had scale:

1. 5 support to 1 customer: 5/1 ratio
2. 6 support to 2 customer: 6/2 ratio
3. 6 support to 3 customer: 2/1 ratio

This type of scaling will hold true til you have a few dozen or hundreds of customers and then something interesting happens. Your number of support staff per customers starts to plummet. Eventually your support staff can handle 100s or 1000s of active customers each. How does this happen?

You beef up your tooling, you provide better documentation, you customers develop super-users, your community develops… your cost of support never stops dropping as a percentage of costs of your software. The marginal cost of customer support approaches zero.

The marginal cost of maintenance & software improvements approaches zero.

Maintenance costs us money. In SaaS, we cannot provide a static product and thus we do not have 0 costs. Again, the statement is the marginal cost of software approaches 0. Providing some scale – each additional customer decreases the relative amount of money it costs to support and improve the product. If you have 5 developers for 5 customers = your ratio is 1/1. As you increase to a 1,000 customers – you will not have 1,000 developers. With a team of 50 – you decrease your ratio to 1/20 devs to customers. The marginal cost of maintenance & software improvements approaches zero.

Most software companies scale their development teams with early revenues or investments because they know when they get close to their “magic size” they can slow their scaling dramatically for future growth, but have the infrastructure to support massive growth.

Competition is a factor in development spend

The realities of the 5 forces (bargaining power of suppliers, threat of substitutes, threat of new entrants, bargaining power of buyers, industry rivalry) demand that SaaS solutions constantly innovate to maintain their position… much less pick up addition market share. Importantly this does not mean you cannot increase your prices, but improvements are usually rolled into the existing product and new modules are usually rolled into a higher tier or new module. What’s interesting about these dynamics is that strength of competition usually helps drive the decision making on what should be rolled in or modularized/tiered. The goal is to have a richer, more capable, product at a lower price than the competition so new customers to the market will have an easy time choosing your solution.

For existing customers in the industry it’s considered much harder. Peter Theil’s rule in Zero to One is that for someone to switch from one product to another for the same price – your solution must be 10x better. There is a race to lower margins as a market is saturated with competition because it becomes extremely difficult to 10x your competitors. The most margin is captured at the beginning of a market when there is low competition.

The marginal cost of infrastructure & hardware approaches zero.

Starting up the cost of hosting is relatively expensive – you have DBAs, Network, Devops – poor tooling and many manual processes. As you scale the amount of staff increases at a slower rate than the rest of your growth – thus causing a decreasing marginal cost. One area that provides for the scaling benefits of SaaS rely on a multi-tenant architecture. In addition, the hardware we use decreases in cost per unit of productivity every year. Thus our newest iterations are faster for less money. The marginal cost of infrastructure goes down over time.

The marginal cost will never be zero

However the software industry shows the benefits of economies of scale dramatically. This is one of the reasons growth is valued over profit in the beginning. Capturing a significant portion of the market allows you to have the scale your competition does not – this is a competitive advantage.

Often when you create a product you know exactly how you would use it, but you have no clue how the end user will use it. Years ago, I ripped out an advertisement from a magazine to remind me of this.

Today, I was reminded of this (and had to dig out this advertisement from hiding) while looking at HubSpot’s Facewall.

You see this was intended to provide you with a way of showcasing internal employees to other employees and add to the sense of team. I immediately thought, I wonder if we should put up dedicated monitors for our enterprise clients and load the top 20/30/50 contacts in those companies in there so everyone can see who we are working with.

Then I got the idea that this would be a really cool solution to tie into a CRM. It’s too bad HubSpot doesn’t have a CRM… oh wait they do.

I was meeting with someone and we were talking about conversion flows and targeted marketing.  I mentioned I always had an idea to target visitors based on if they had visited a competitor.

Black hat CSS

So this guy {rename nameless to protect the guilty} – says I looked into it and the browsers disabled the :visited css hack you were thinking about using.

I bounced a couple ideas around in my head and came up with this.

Black hat Javascript

Let’s see if this theoretical workaround would work:

1. Browser: Set Visitor ID

   visitor-id: 123456

2. Browser: Create the CSS for the Visitor

   .competitor-1, 
   .competitor-1:visited 
     {position:absolute; top:0px; right:0px;}
   
   .competitor-1
     {background-image: black-pixel.jpg
      ?visitor-id=123456&competitor-1;}
                            
   .competitor-1:visited  
     {background-image: white-pixel.jpg
      ?visitor-id=123456&competitor-1;}
   

3. Browser: Create the Link with JS
4. Browser: Apply the Styles with JS
5. Server: Read the Meta Data Collected from Image Call
6. Server: Add that Meta Data to User Profile (preferably something like the leadin profile)
7. Both: Cycle Through Other Competitors
8. Server: Serve Marketing Content
9. Server: Check to see if the user has left the site and come back – if so re-run and then
   tag if competitor has visited the site (now you can add within a specific time period)

Targeted Customization

So what can this help with?

• Calling out specific features/functions/pricing relative to a competitor
• This can help guide c2a, popups, on-page content, & even pricing (for instance people on a Mac get charged more on some e-commerce sites)

Competitor Page

But, I think it could make for a really compelling competitor analysis page. Only show the competitor you know they have visited already : ) and dynamically change the pricing according to the features/funtions/pricing you have for those competitors.

Disclaimer: These are just thoughts…

Ok – I must admit I enjoy thinking about the dark arts of marketing sometimes. This one is mostly a puzzle challenger and not something I endorse (publicly).  Certainly not something, I’m technical enough to do, but thinking about the ability to automatically add tags for all the competitors to the leadin/crm profile of your users – should make marketers weep tears.

Does anyone see a reason this wouldn’t work?