Adventures in AI economics; SaaS+fintech; growth+sales; and GPT-3’s prospects

IN THIS EDITION

• Adventures in AI economics
• When SaaS adds fintech: bigger markets & better margins
• The hidden traps of “growth+sales” – and how to avoid them
• The tantalizing prospects of GPT-3

Adventures in AI economics

Matt Bornstein, Martin Casado

More companies are incorporating AI/ML into their products to deliver core functionality, but as we’ve shared before, the economics of AI businesses (as compared to software businesses) are HARD. This is because AI development is a process of experimenting and taming the complexity of the real world, much like physics or other physical sciences. The data is often messy and full of edge cases, resulting in long tailed distributions.

So, given the long tail of AI – and the work that it creates – how do you improve the economics of an AI business? We talked with dozens of leading ML teams, and addressing the long tail starts with understanding the distribution of the problem you’re solving.

1. Not actually a long tail (easy): if you have a well-bounded problem, you likely do not need ML or AI. Logistic regression and random forests are popular for a reason, so start simple and upgrade to bigger models only when the problem demands

2. Global long tail (hard): if you have a similar distribution for all customers, you can optimize, narrow, reframe, and explore a growing technique called componentizing, where a single problem is broken into smaller pieces

3. Local long tail (harder): when the distribution of data looks different for every customer, the strategies for handling it are still nascent. Meta models and transfer learning may be able to help, although we haven’t yet found examples of them being used successfully at scale, and trunk models seem like an emerging API standard for ML.

Read our post on AI/ML techniques and strategies »

When SaaS adds fintech: bigger markets & better margins

Kristina Shen, Kimberly Tan

Today, the majority of SaaS revenue comes from subscription fees, but new infrastructure players have made it easier and cheaper to add fintech. The next evolution of the software business model is a hybrid of subscription SaaS plus fintech, and we’re seeing this evolution happen first in vertical SaaS.

Vertical SaaS customers tend to prefer purpose-built software for their specific industry and use case, often leading to a single dominant solution in a particular vertical and winner-take-most market dynamics. By adding financial products and services, vertical markets become even larger by increasing revenue per customer.

While early fintech in SaaS has been mostly reselling payments for a referral fee, there is increasingly an opportunity to embed fintech into SaaS products. Embedding fintech adds cost and complexity, but has the potential to repay with bigger margins and a better customer experience. And as more SaaS companies add fintech, we expect to see more go beyond payments into lending, cards, insurance, and other financial services.

Read our full breakdown of fintech models for SaaS »

Coffee-break worthy: Taylor Swift as B2B Software and D&D with a GPT-3-powered AI

The hidden traps of “growth+sales” – and how to avoid them

Peter Lauten, Martin Casado

“Growth+sales” – where organic adoption by users comes first and traditional top-down selling comes later – has become the predominant go-to-market (GTM) model for enterprise software. In growth+sales, all the complexity of the traditional sales cycle collapses into a single decision maker: the end user. This has made enterprise software more user-friendly, shifted the focus from pre-sales to post-sales, and perhaps, most importantly, given the advantage to startups and product visionaries.

Still, the growth+sales GTM model comes with its own traps as companies scale, from the pre-product market fit days to the later stages – not least because it’s hard enough to get one GTM motion right, much less two at once. Based on what we’ve seen in growth+sales companies, these are some of the common failure modes – and the questions and tactics that can help teams avoid them.

Failure Mode #1: Forcing the growth motion when it doesn’t fit the product
Before even starting down the growth+sales journey, make sure it’s right for your startup. Is the out-of-the-box value proposition to the end user sufficiently compelling for them to adopt on their own? Can end users adopt without a heavy-touch implementation process that requires engineering resources to integrate?

Failure Mode #2: Buying end user mindshare that doesn’t spark organic love for the product
Do your end users truly love your product? If organic adoption is the holy grail for bottom-up adoption, then paid acquisition channels that yield low-retaining users is the addictive drug that will ultimately kill a distribution model.

Failure Mode #3: Using traditional metrics to measure a new motion
Traditional sales metrics (e.g. ARR, ASPS, etc) aren’t sufficient to understand how well the growth motion is working. Instead, look to consumer-style metrics that track active user growth (MAUs, WAUs, DAUs), engagement (session length, days active in a given work week), active user retention, and funnel conversion points throughout the user journey.

Failure Mode #4: Layering on sales too early
Before layering on top-down sales, look for evidence from the market, as users ask: “How can I get this in the hands of my entire department?” “How do we unlock all the premium features?” “Can you talk to our IT department to get this rolled out to my entire department?” “Why isn’t everybody using this?!”

Failure Mode #5: Growth momentum that doesn’t actually make enterprise sales any easier
Does the growth motion lead you to the right buyer? Or are users perfectly happy with the freemium product?

Failure Mode #6: Neglecting the growth motion once sales takes off
Growth+sales is not about doing growth first, and then growing up to sales without ever looking back. It’s about delivering on your promises to the user, then making new promises to the enterprise buyer without forgetting about the user.

Read “Growth+Sales: The New Era of Enterprise Go-to-Market” »

The tantalizing prospects of GPT-3

There’s been a lot of excitement and viral sharing of examples around GPT-3 and the commercial API from OpenAI – the pre-trained machine learning model that can perform a variety of natural-language processing tasks – but what are the implications for startups and big companies, and really, any product built with “AI inside”?

The tantalizing prospect is that any startup that wants to solve a natural language problem – like chat bots for pre- or post-sales support, summarizing documents, or even just going through every customer complaint ever and finding insights for product managers – could quickly and cheaply build on top of that infrastructure.

However, it’s still only a tantalizing prospect, because as with the Greek myth of Tantalus, we’re still reaching for the promise of this. There’s still a huge gap between other very straightforward and usable APIs and Open AI’s API. The hope, however, is that this or something like it could dramatically reduce the data gathering/cleansing/cleaning process – and even building the machine learning model – dramatically reducing the time it takes to build a “machine-learning inside” product. But it’s still too early to tell.

Listen to our discussion of all this on 16 Minutes »

---

What did you think of this newsletter? Let us know at enterprisenewsletter@a16z.com

image: David Traña | Unsplash