Key Aspects of an Enterprise Data Platform
1. Transforming Workflows and High Engagement.
Disruptive technology platforms become an essential part of an individual’s daily workflow and cannot be easily replaced. The product’s benefits must be transformative, not incremental (i.e. minimum of several hours saved per week per user, or gross margin improvement of 15% or greater). The best way to track success here is to closely follow product engagement metrics, such as number of visits per user or average time spent per week using the product. Having this kind of direct connection also enables platforms to turn their users into a valuable asset for third party developers. Traditional enterprise middle-ware and backend solutions don’t have these innate features and should be avoided.
2. Generate and Capture Proprietary Data.
In the Smart Enterprise world, Data Is King. Having proprietary data that is user-generated and owned by the platform is becoming a pre-requisite to building a defensible business. LinkedIn, for instance, encourages users to publish content about themselves that cannot otherwise be found online. This pulls other professionals (and recruiters) onto the platform. The data is also highly structured, giving it valuable second order applications, such as targeted advertising. The most successful platforms will be open to third party development, but in order to be valuable, consumers and businesses must be forced to come to you for data that they cannot get elsewhere.
3. Network Effects and Infrastructure.
The value of the platform (to the end user) should grow with the number of users in a virtuous cycle that encourages adoption virality. Ebay, Microsoft Office, and Visa are all classic examples of this phenomenon. As a general rule, a system that controls the reciprocal exchange of information, goods, or money will tend to gain network effects. The largest and most enduring Smart Enterprise platforms will likely have B2B marketplace and communication/collaboration features at their core. Enterprise applications that leverage data and user engagement to act as infrastructure for a robust third party development ecosystem will also tend towards winner-take-all dynamics, as occurred with the Microsoft Windows operating system.
4. Owning the End Consumer.
The trend of the “consumerization of the enterprise” is not just about designing better user interfaces for employees; it involves engaging consumers directly, since they are a valuable source of data for companies. This phenomenon will become particularly apparent in Healthcare, Finance, and Government. Smart Enterprise software will change the dynamics of enterprise-consumer interaction by bringing the two parties onto the same data platform. For instance, consumers might use an online tool to track their health information and set fitness goals. Physicians could then utilize this data to monitor their patients’ ongoing health, while insurance companies are able to reduce claims by offering premium discounts to consumers who exercise regularly.
5. Building a Brand.
Winner-take-all-players often become brands that perpetuate their market position. This becomes especially relevant once you move outside of the early adopter subset and into the mass market. Professionals (especially in Medicine and Law) often place a high degree of importance on a product solution’s status as “the gold-standard”, or trusted authority. There will undoubtedly be Smart Enterprise content companies that leverage new analytical and data collection techniques (such as using mobile devices to crowd-source consumer good pricing data in the developing world) to build the next-generation of defensible content businesses to compete with existing brands like Nielson and Thompson Reuters.
Common Pitfalls & Signs of an App, Not a Platform
1. Sales vs. Product-Driven Leadership.
Successful technology platforms require a degree of long-term planning and vision that sales organizations are not well suited to deliver. Sales teams can be great at building a market for a new technology, but as customer needs evolve beyond linear software tools and SaaS pricing models shorten commitment periods, product and data strategy will become a critical part of gaining lock-in and winning a space. Box, for instance, is using a sales team to gain ground in enterprises, with the intent of scaling back these efforts once it achieves critical mass in specific industry verticals. LinkedIn also relied on an early sales team to jump-start engagement from recruiters and companies, which increased adoption by professionals and strengthened the data platform.
2. Product Only Ingests Public Data.
The novel acquisition of public data can be an effective way to create an initial value proposition, but it is not defensible over the long term. We have seen many enterprise applications pull in data from social networks like LinkedIn and Twitter and then run analytics on it — “machine learning”, “semantic analysis”, and “natural language processing” are the new buzz words of this current technology wave. This might be a “Big Data” play, but the critical question that must be asked is where is the defensibility? If the product is simply pulling in the same data that anyone else can extract via public APIs, what stops another team from doing the same thing incrementally better?
3. Reliance on Technology Over Strategy.
In today’s competitive environment, technological prowess is a pre-requisite to business success. However, technology is not a defensible attribute over the long-term. Technology (such as a more intelligent data parsing or prediction algorithm) can help create an initial value proposition, but given the current velocity of innovation, the competition is never that far behind. Software patents don’t hurt, but they shouldn’t be considered core to the business model. Ultimately, unless you capture valuable proprietary data or become an irreplaceable part of the customer’s workflow, you have not achieved defensibility. Historically, the most successful companies — like Google or Apple — use technology to create disruptive value but then quickly build a defensible platform and network around it to protect their position (such as app ecosystems reliant on the platform’s proprietary data).
4. Claims That “Data Scale” Is a Platform Effect.
Many companies — especially tools that rely on prediction and recommendation algorithms — will claim that by increasing the volume of data being captured by their application, they will be able to use machine learning to improve product quality and gain an edge over the competition. Scaling data is certainly important, but once you hit the point of statistical significance we believe the advantage tends to wear off and you need other elements of defensibility — like proprietary data or network effects — in order to win a market. The exception may be in networks where a high degree of value is created by outliers (such as in scientific or medical research communities) and it is important to capture the long tail.
General Partner, 8VC