The Product Discovery Model

(There’s a presentation version of this post here.)

The Product Discovery Model is an alternative to the Internal Startup Model for inventing digital consumer products, designed to address many of its shortcomings. The Product Discovery Model reframes the tactical purpose of the innovation work from building new products to reducing uncertainty about product/market fit. The shift in perspective is not new: Eric Reis popularized the paradigm shift from building to learning in his book Lean Startup, which has become a mass movement in product management. If you are familiar with his approach, you can think of the Product Discovery Model as Lean Startup tailored for the risk/return portfolio mentality of corporate investments. Startups generally think of innovation as a do-or-die pursuit of any kind of product/market fit within reach, whereas corporations need to think of innovation as a calculated way to change the course of the business through a process of optimizing risk/return in a curated portfolio of product options. Each product option has the potential to transport the business somewhere it wants to go that the normal, rational process of product optimization cannot reach quickly or incrementally.

As uncertainty in one innovation idea is reduced through a process of validation, the evaluation of the product option is updated in an iterative process until we are comfortable with either discarding it or investing in it. You can think of the process as a matrix with a classical product funnel as rows, and a discovery funnel as columns. Certainty increases from top left to bottom right.

Another way to think of it is that the Product Discovery Model takes the stance that it’s best to assume that the Trough of Sorrow is inevitable and lethal, and thus every new product’s immediate objective is how to shorten its sojourn there. (The Trough of Sorrow, a term introduced by Paul Graham as a joke at dinner, is that floundering period in every innovation’s life where plan meets elusive reality; very few products ever make it through and, because of the culture clash of innovation, it’s particularly deadly in traditional corporations.)

Always Be Pitching

The Product Discovery Model encourages us to imagine the product as an ideal Product Pitch that’s updated like a journal. The Product Pitch is a complete summary of the product and its market fit, with each proposition tagged for level of certainty. Each step going from left to right and top to bottom results in an update to the Product Pitch, making it more and more accurate and reliable, up until the point where we decide to discard the idea or invest it in fully.

Unlike a traditional product or business plan, the Product Pitch is meant to be aspirational and involve some hand-waving at the start of the process; too much initial confidence and precision is suspicious, unless it is backed up by evidence from a very similar product (as might be the case for a product clone). Progress is measured by how much ‘validated learning’ (to use Eric Reis’s phrase) is poured into the Product Pitch as you work your way down the product funnel.

Product Pitch Components

1. Personas: Who for (persona, market segment), how many potential users, and how addressable? (Total Addressable Market or TAM)
2. Benefits: What are the product benefits and how important to users? (the effect of solving the problem / the job to be done / the progress to be made)
3. Features: How are the benefits achieved? (description of the solution, such that both a lay person and engineer could understand how it produces the effect that generates the benefits)
4. Positioning: What is the competitive position? (value relative to all relevant alternatives, including people who don’t use the product category — the coveted non-consumers category)
5. Resources: Level of effort to get to market with an MVP? (resources, time, money)
6. Marketing: How are new users acquired, for what cost?
7. Retention: How are users retained, through what reactivation mechanisms and marketing channels?
8. Revenue: How is revenue generated?
9. Overall Value — state full value proposition a) for users & b) your business (ARPU, ROI)

The Minimum Viable Product

A note on the complexity of Innovation products. Innovation products should always be “Minimum Viable Products” (MVPs). This cannot be emphasized enough, but it is easier said than done. Think back to the early versions of entrants in almost any new product category and consider their level of complexity: Facebook was a way to find out who you are connected to on campus; Instagram was a camera photo filters app; Snapchat was a disappearing photo messaging app; Google was a search engine based on the conceptually simple Page Rank algorithm. Nearly all successful innovative apps start life with a very simple focus on one thing and evolve from there. Simplicity is what made it easy for them to gain adoption and evolve rapidly. Most of the apps mentioned above are now so complex and presuppose such vast networks of user adoption that if they were launched now in their current relatively bloated state, they would almost certainly fail.

But it’s not so easy to figure out the minima – it’s the kind of thing that will seem obvious to everyone in retrospect but will involve explorations down blind alleys chasing concepts that in retrospect will seem stupid. If the product is a success, this period in its history will be quietly rewritten to conform better to a hero genius myth of divine inspiration (does anyone remember or care that Twitter or Slack were accidental by-products of official failures?) This process is almost never appreciated (it’s very meta, hardly perceived as a process at all, more like accidental history or fate). But it is an essential ingredient in the product discovery process whether we see it or not. An advantage of the Product Discovery Model is that it makes this history explicit.

The Product Funnel & the Importance of Sequence

The Product Funnel is the causal sequence of outputs that make up the functioning of a successful digital consumer product.

Product/Solution Fit: people want it and it works
1. Attraction: appeal of value proposition, brand messaging, users taking an interest

2. Acquisition: effectively get users to install & register

3. Activation: getting hooked on something in the first 1–24 hrs

4. Engagement: completing feature funnel conversions

Product/Market Fit: predicted positive ROI on growth
5. Reactivation: return loops

6. Referral: viral loops

7. Revenue: monetary pinches

Because the outputs are sequential, they also determine the order in which product innovation challenges should be tackled. This last point is critical, but it is not self-explanatory.

It’s natural to think of an idea for a product as a solution to some problem, that an inventor builds and tests before figuring out how to attract users and market it. Lots of products are launched this way, but the likelihood of any one consumer product succeeding this way is essentially like betting the farm that you figured out the right configuration of a hundred product, marketing, and economic inter-dependent variables in a single shot. Why is this? Consumer needs are very open-ended (compare the desire to ‘feel wanted’ to the more practical desire to mend your shoes), discovering new perceived needs is hard — often going hand-in-hand with creating the perception of those needs and new behaviors around them (cf fashion) — in a market that is very competitive, changeable, and uncontrollable. If your app is social, you have an additional layer of wicked cold-start problems whose solution requires a critical mass of network effects, mutual reciprocity, and social density before the product is even useful to its first user. If the market is social and competitive, you have winner-take-all effects to contend with. If your market is teen or pre-teen and social, you also the problem (also an opportunity) of social cultural plasticity — sometimes known in the evolutionary psychology field as neoteny — that is, the almost endless ways humans can intermediate their social identity, relationships, and groupings with cultural technology.

Faced with this uncertainty, if you want to manage your innovation risk you must start at the front of the product funnel’s causal chain, even if that means creating ersatz experiences to stand-in for the substance of your product down the line; even if that means the substance of your product idea paradoxically changes as you find the pathways to get traction.

This is also the reason that corporations are nearly always better off leveraging the products and audiences they already have than to go after completely new markets from scratch. The further up the ladder you can start, the multiplicatively better chance at succeeding. Caveat: the higher up the ladder you start, the more you will be locked into whatever unforeseeable limits your business currently has (for many over-the-hill businesses these limits are severe). That is why it is good to have a risk-adjusted portfolio that includes some cheap product options with massive potential upsides outside of your core business; along with larger initiatives with much higher probabilities for success but smaller ultimate upside.

The Discovery Funnel & Scientific Method

Returning to the matrix described at the head of this article: each step of the Product Funnel goes through the Discovery Funnel. The discovery funnel (Distill, Validate, Prove, and Invest) is essentially the scientific method. More on the discovery funnel in a future article.

Next Steps

For the cultural challenges of using this model in practice inside a traditional corporation, see the first article in this series: The Culture Clash of Innovation.

(There’s a presentation version of this post here.)