written by
Bret Carmichael

What big companies need to know about software innovation

Innovation 13 min read
Space shuttle, as software innovation, taking-off in a new direction

Large companies often regard software innovation programs with a healthy amount of skepticism. Executives think of them as a pit, into which they throw money never to be seen again. Such programs offer an opaque view of resource outputs and are at best, an HR nicety implemented to attract and retain software developers during a time of resource scarcity. The prevailing view is; innovation programs are about marketing. There's little value gained by investing money in something with unexplainable benefits. However, well executed software innovation programs are more than a nicety. They serve a critical need; to answer external market challenges enabled through software. Here, we'll pivot both the positioning of software innovation programs within companies and how they need to be viewed by leadership.

Democratization of capability

The cloud provides several advantages over legacy client-server architectures. The management of infrastructure as code enables quick build times and drastically lower run costs, so low that Amazon Web Services (AWS) offers a free tier to entice new customers. This has fundamentally changed how teams operate. It's also changed how IT business operations teams function.

Resource acquisition costs

Hardware acquisition was formerly budgeted. When a new server needed to be purchased, there was a specific decision point that needed to occur. Someone had to sign-off on the purchase of the server. A purchase order was issued, assuming the department's current quarter or annual performance to budget was on track, which was not always the case. When a postponement decision was made, it was frequently accompanied by a delay in software delivery. Conversely, now, any employee can build a server and deploy an application within an hour. Cost pressure is nonexistent. The company pays a predictable service fee. Loaded in that fee is tolerance for building and destroying software infrastructure components at high frequency. Related asset management functions are all but eliminated from the process of delivering infrastructure.

Lower costs are a double-edged sword for large companies. Saving money while gaining increased scale, speed, and reliability feels good. It's measurable. It's a great story, and it's why a lot of companies are aggressively working to shift their workloads from on-premises to the cloud. The democratization of capability, however, introduces new risks.

Increased competition

The other edge of the sword is; new competitors don't have to shift any workloads. They can start in the cloud and practice spend avoidance, while companies with legacy, on-premise application stacks need to invest time and money in migration efforts. Further, it must be noted that, while shifting workloads is a mighty achievement, it's a function of operations: a large maintenance project. This is not to say that, maintenance isn't important. It is. Rather, the point is made to highlight the fact that, large cloud migration projects coincide with gaps in feature delivery. Their outcomes are looked upon with admiration internally, but occupy resources with work that is invisible to the user.

Competition exists in higher numbers. With cost no longer a barrier to market entry, enterprises must compete with garage startups. Anyone with an idea and the necessary skills can enter a market and dynamically scale to meet demand. This puts large companies at a distinct disadvantage due to carried financial overhead intrinsic to affected markets. It is more expensive for the enterprise to enable a given capability than it is a new market entrant, and the market entrant is happy to offer a lower price. The enterprise is faced with a difficult decision; to cannibalize an existing revenue stream or lose the market. Both choices mean revenue contraction and contradict the behaviors that made them successful.

Disruptive effects of software innovation


Amazon and its app disrupted retail through software.

It's no secret, software has had a transformative impact on multiple industries. Amazon's displacement of brick and mortar retail was enabled by software. Formerly, a person would drive to a store where they could touch and feel products and get expert help from staff. Gratification was instantaneous. They left the store with their item. Amazon provided an alternative: convenience and cost savings. Customers could no longer physically inspect goods before purchase. They lost access to expertise and had to wait seven days to received their item, but shopping online was convenient, and it was cheaper. It was good enough.

With Amazon's growth, they introduced new features. For $80 a year, customers gained access to two-day shipping through Amazon Prime. They received their items faster: a more premium and upscale experience. Then, new features were added to Amazon Prime, like Prime Video and Prime Music. These features weren't related to retail or the sale of physical goods. However, because they made Amazon Prime more valuable, the service moved upmarket, sustaining multiple price increases; to $99 and $119 a year.

With traditional brick and mortar retail in a weakened state, Amazon began building brick and mortar retail establishments. They entered the very same space they had disrupted through software.


Netflix disrupted software through software.

Netflix has one of the more famous disruption stories. Through software innovation, they disrupted video rental chain, Blockbuster.

At its peak, Blockbuster had more than 9,000 retail stores. Customers would drive to one of their stores, rent a film for one or two nights, and return it at the end of the rental period. They enjoyed near-immediate gratification; getting to watch their chosen film as soon as they arrived home. However, renting movies was expensive, and high late fees were a point of contention with customers.

Netflix entered the DVD rental business with no retail stores. Instead, they used software and the internet. Customers would pay a monthly fee. In return for that fee, they would select a DVD, and Netflix would send it to them in the mail. The customer needed to wait days to receive the DVD, but they could keep it as long as they wanted. Alternatively, they could return the DVD soon after watching it and select another one. Netflix customers didn't have the benefit of instant gratification, but Netflix was cheap and void of late fees.

Netflix later introduced streaming, increasing the value of their service. This allowed them to charge more and go upmarket. This was followed by original content, which allowed them to charge more, compete with move theaters, and go further upmarket. Netflix subsequently added UHD (4k resolution) and HDR (High Dynamic Range) content. Again, they were able to charge more and move upmarket.

Netflix still offers DVD rentals, but the lower cost DVD rental business has been all but entirely supplanted by more its expensive streaming service. Streaming provides instant gratification, more so than Blockbuster had provided. Customers don't need to drive to a store. Instead, they open the Netflix app and choose what they want to watch, unconstrained by availability.

Barriers to software innovation in the enterprise

Given a multitude of outcomes, only a small percentage are relevant. Marc Andreessen, founder of Netscape and venture capital firm Andreessen Horowitz explained this effectively.

"The key characteristic of venture capital is that returns are a power-law distribution. So, the basic math component is that there are about 4,000 startups a year that are founded in the technology industry which would like to raise venture capital and we can invest in about 20...We see about 3,000 inbound referred opportunities per year we narrow that down to a couple hundred that are taken particularly seriously… There are about 200 of these startups a year that are fundable by top VCs. … about 15 of those will generate 95% of all the economic returns … even the top VCs write off half their deals." - Marc Andreessen

Power-law distribution

As the volume of product/market entrants increases, so too does the likelihood of one being successful and paying for the others that fail. This is important to note, because it suggests how investors can become hyper-selective. It also requires highlighting role contention between the traditional CIO and the CTO.

Role contention between the CIO and CTO

Definitions can vary; at a high level, the CIO is a function of operations. CIOs are beholden to an organization's directed business goals, in the specific context of its market, and use technology to capture cost savings and extract more value from customers. By contrast, a CTO is designated with the task of extending internally developed software externally -- as its own product(s) -- and growing it. There's contention when an organization has only a CIO, or the CTO works for the CIO and is a discretionary line item – contingent to goals established elsewhere in the company. Innovation is the first budget cut.


Enterprises have developers and development teams. Despite being named as such, IT organizations are often entirely a function of operations. They're cost centers, not profit centers. Companies invest in profit centers and contain cost centers. When considering innovative product areas, this creates an environment of hyper-selectiveness, one that is not conducive to software innovation.

"How will we make money off this?" may be the right question at a time in the future, but it's nearly impossible to answer when exploring the merits of a product idea that serves a new product category. The answers can only be, "I don't know," or a confident, yet unsubstantiated explanation of a business model. Both answers are harmful. The first sets the stage for a blanket denial. The second is a path to disappointment, when the product is expected to succeed, but results in probable failure. Outside the scope of this article, there are much better questions to ask early when evaluating the merits of a new product category.

Operational IT organizations will invest in product categories, when the likelihood of success has a high degree of certainty. This yields multiple effects:

  • Infrequent investments
  • Investments that are limited to sustaining innovation, with the organization's market
  • Low tolerance for product categories that will not achieve a short-term return on investment
  • High barrier for interest and accumulation of dismissals; a perception that the software innovation program is meritless

Understanding the core challenge

Innovation is not an event. All successful software innovation programs require pipelines that continuously introduce and consider a multitude of product ideas in a diverse set of product categories. From the multitude of ideas, products must be brought to market with the expectation that most products brought to market will fail. Marc Andreessen explains it this way.

“We think you can draw a 2×2 matrix for venture capital…And on one axis you could say, consensus versus non-consensus. And on the other axis you can say, successful or failure. And of course, you make all your money on successful and non-consensus...It’s very hard to make money on successful and consensus. Because if something is already consensus then money will have already flooded in and the profit opportunity is gone. And so by definition in venture capital, if you are doing it right, you are continuously investing in things that are non-consensus at the time of investment. And let me translate ‘non-consensus’: in sort of practical terms, it translates to crazy. You are investing in things that look like they are just nuts.” - Marc Andreessen

For operational IT organizations, Andreessen's line of thinking goes against the grain. It's true of most investments; CIOs are unable explain how any single investment in a new product category will contribute to financial targets in the current year or five years in the future. By eliminating related spend, the company can improve profitability.

This is a problem for companies operating in staid markets, as yet undisturbed by software innovation. It relies on the hope that a competitive upstart won't bring a disruptive product idea to market and leaves large companies at risk for rapidly eroding revenue streams and struggling to find opportunities for meaningful recovery.

One or two safe product investments cannot save a large company from decline when faced with the evaporation of its market. Further, the safe and profitable investments may be subject to the same decline when existing in the same market.

Enterprises have human scale

Large companies have an advantage of scale. However, that scale is untied to their ability to provision technology. That capability is fully democratized. The enterprise's scale is in its employee numbers and creative talent. It's also well funded. Thus, large organizations have access to and can hire the best talent. Absent venture capital funding or promise of equity, small companies are unable to compete.

Through scale of individuals and talent, enterprises can cultivate and develop robust product innovation pipelines. Moreover, participation in such pipelines is not limited to IT employees with a software developer skillset. The more employees a company has, the larger its innovation pipeline can be.

Large companies have capital and capacity to allocate a percentage of their budget to investment in new product areas. However, for this to occur, software innovation centers must be regarded as profit centers and budgeted as such. And, it will take some time for accrued gains to make the program profitable.

Access to innovation

Robot posing for picture

There are two ways companies can innovate. They can invent, leveraging employee talent to create new products and bring them to market. Slack is one such example, first developed as an internal communication tool by game developer Tiny Speck. Twitter, another example, was created as a side project of podcasting company Odeo. Both progenitors are gone. Their offspring, now household names, bear no resemblance to the companies from which they were grown.

The second way to innovate is through acquisition. Companies can buy software innovation. Famously, Facebook's acquisition of Instagram was one of the most successful purchases in recent memory. YouTube was a similarly successful purchase for Google after their own attempt at video failed.

Investment in internal capital has a longer tail, but carries a lower cost. Conversely, the benefits of acquisition are often felt more quickly.

Purchasing innovation is considerably more expensive than cultivating it, and may not reduce the risk of market displacement. Large companies are unlikely to invest large sums of money in markets outside their niche expertise: product areas they don't fully understand. This makes practical sense, but means acquisitions mitigate less risk than do innovation programs in staid markets. Should a maker of fax machines purchase another maker of fax machines, it will still be true that, use of fax machines is in a state of permanent decline.

Diversification of risk in the face of multiple threats

Amazon disrupted and owns 6% of all retail. They're vertically integrating, with their own fleet of cargo planes and delivery trucks. They operate a thriving business in Amazon Web Services, create and distribute content through Prime Video and Prime Music, and have a suite of Echo products. They also own Whole Foods' grocery business and Eero's networking business. Amazon is diversified.

Google has a plethora of well known internet software products, like Gmail. Like Amazon, they're diversified and integrated in several other product areas, including:

  • Google Ads
  • Google Cloud
  • Google Ventures
  • Nest
  • Owned Boston Dynamics (a robotics company)
  • Google Home
  • Google Ventures
  • Waymo (self-driving cars)
  • G Suite
  • Chrome OS
  • Chrome Browser
  • Android

Neither Google's nor Amazon's success has come without failure. Amazon's Fire Phone was an expensive one, but it led to Echo and features like product scanning in their shopping app. Google is famous for sunsetting products and services. A website exists to catalog them, Google Graveyard, with more than 150 listed.

Large companies can innovate

The enterprise can experiment and innovate just as tech companies do. The difference is company culture and positioning. All companies use technology. What differentiates a tech company from a niche company is its willingness to enter new markets and sustain failure therein.

  • One VC firm funds individual startups with uncertain outcomes: high risk, high potential
  • Many startups self-fund individual investments with uncertain outcomes: high risk, high potential
  • One large, niche company funds few internal capital investments with more certain outcomes: low risk, low potential

Democratization of scale has lowered the barrier for market entry and increased the volume of competitors. Large companies need healthy software innovation programs to offset persistent market threats. Such programs need product diversity; to fearlessly explore opportunities in different markets, because the current market might only exist at a fraction of its current value tomorrow.