Simply put, a product or service has a network effect when it becomes more valuable as more people use it/ devices join it (think of examples like the telephone network, Ethernet, eBay, and Facebook). By increasing engagement and higher margins,network effects are key in helping software companies build a durable moat that insulates them from competition.

However, there is no single metric to demonstrate that a business has “network effects” (Metcalfe’s Law is a descriptive formulation, not a measure). But we often see entrepreneurs assert that their business has network effects without providing any supporting evidence. It’s hard for us to resolve whether a business indeed has network effects without this — leading us to more heated debates internally as well!

Let’s use OpenTable as an example of a business with network effects. The OpenTable network effect was that more restaurant selection attracted diners, and more diners attracted restaurants. Here are some of the measures that helped demonstrate those network effects (we typically used measurements within one city to illustrate the point, as OpenTable’s network effect was largely local):

The sales productivity of OpenTable sales representatives grows substantially over time, due in part to large increases in the number of inbound leads from restaurants over time.  This is more meaningful than the fact that the total restaurant base grows over time, as that can happen even without network effects.

The number of diners seated at existing OpenTable restaurants grows substantially over time. This again is more meaningful than the fact that the total number of diners grows over time.

The share of diners who come directly to OpenTable to make their reservation (versus going to the restaurants’ websites) grows substantially over time.

Restaurant churn declines over time.

As you can see, most of these metrics are specific to the network that OpenTable is building.  Other network-effects businesses — such as Airbnb, eBay, Facebook, PayPal — have very different metrics.

So the most important thing in managing a business with network effects is to define what those metrics are, and track them over time. This may seem obvious, but the more intentional you are about — vs. “surprised” by — your network effects, the better your business will be able to sustain and grow them. Similarly, it’s important for prospective investors to see evidence of a network effect, that the entrepreneur understands exactly what it is, and how he or she is driving it.

Sell-through rate is typically calculated in one way — number of units sold in a period divided by the number of items at the beginning of the period — but has different uses and implications in different types of businesses.

In marketplace businesses, sell-through rate can also go by “close rate”, “conversion rate”, and “success rate”. Regardless of what it’s called, sell-through rate is one of the single most important metrics in a marketplace business. As investors, we like to see a relatively high rate so that suppliers are seeing good returns on the effort they put into posting listings on the marketplace. We also like to see this ratio improving over time, particularly in the early stages of marketplace development (as it often indicates developing network effects).

In businesses that buy any kind of inventory — retailers, wholesalers, manufacturers — the sell-through rate is a key operating metric for managing inventory on a weekly or daily basis. It can reveal how well you matched supply of your product to demand for it, on a product-by-product basis.

For many investors, however, inventory turns is a more useful metric than sell-through rate in inventory-based businesses, because it:

— Talks to the capital efficiency of the business, where more turns are better

— Provides clues as to the quality of the inventory, where slowing inventory turns over time can signal slowing demand as well as potential inventory impairments (which can lead to mark-downs or write-offs)

Inventory turns typically are calculated by dividing the cost of goods sold for a period against the average inventory for that period. The most typical period used is annual.

There are two different ways to improve inventory turns — (1) By increasing sales velocity on the same amount of inventory; (2) By decreasing the inventory needed to generate a given amount of sales. While both are fine, one caution on the latter: Managing inventory too closely can potentially impact sales negatively by not having enough stock to fulfill consumer demand.

Gross margin — which is a company’s total sales revenue minus cost of goods sold — can be considered an equalizer across businesses with different business models, where comparing relative revenue would otherwise be somewhat meaningless. Gross margin tells the investor how much money the company has to cover its operating expenses and (hopefully!) drop to the bottom line as profitability. 

A few examples to illustrate the point: E-commerce businesses typically have relatively low gross margins, as best exemplified by Amazon and its 27% figure. By contrast, most marketplaces (note here the distinction between e-commerce) and software companies should be high gross-margin businesses.

Paraphrasing Jim Barksdale (the celebrated COO of Fedex, CEO of McCaw Cellular, and CEO of Netscape), “Here’s the magical thing about software: software is something I have, I can sell it to you, and after that, I still have it.” Because of this magical property, software companies should have very high gross margins, in the 80%-90% range. Smaller software companies might start with lower gross margins as they provision more capacity than they need, but these days with pay-as-you-go public cloud services, the need for small companies to buy and operate expensive gear has vanished, so even early stage companies can start out of the gate with relatively high gross margins.

ARPU is defined as total revenue divided by the number of users for a specific time period,  typically over a month, quarter, or year. This is a meaningful metric as it demonstrates the value of users on your platform, regardless of whether those users buy subscriptions (such as telecom monthly subscriptions) or click on ads as they consume content.

For pre-revenue companies, investors will often compare the prospects of a company against the known ARPU for established companies. For example, we know that Facebook generated $9.30 ARPU in FY2015Q2 from its U.S. and Canada users:

So if we’re evaluating a company with an advertising business that has monetization potential comparable to Facebook, we ask: Do we believe the company can generate a quarter, half, just as much, or even more ARPU compared to Facebook? What would need to be true to justify this belief? How would the company achieve that (and do they have the ability do so)?

While we’ve already made this point in part one of this post, we want to emphasize again that when software businesses use ARR, they mean annual recurring revenue,NOT annual run rate. It’s a mistake to multiply the recognized bookings — and in some cases revenue — in a given month by 12 (thus “annualizing it”) and call that number ARR.

In a SaaS business, ARR is the measure of recurring revenue on an annual basis. It should exclude one-time fees, professional service fees, and any variable usage fees. This is important because in a given month you may recognize more revenue as a result of invoicing one-time services or support, and multiplying that number by 12 could significantly overstate your true ARR potential.

In marketplace businesses — which are more transaction-based and typically do not have contracts — we look at current revenue run rates, by annualizing the GMV or revenue metric for the most recent month or quarter.

One mistake we frequently see is marketplace GMV being referred to as “revenue”, which can overstate the size of the business meaningfully. GMV typically reflects what consumers are spending on the site, whereas revenue is the portion of GMV that the marketplace takes (“the take”) for providing their service.

TAM is a way to quantify the market size/ opportunity. But using the size of an existing market might actually understate the opportunity of new business models: For example, SaaS relative to on-premise enterprise software may have much lower average revenue per user but more than make up for it by expanding the number of users, thus growing the market. Or, something that provides an order of magnitude better functionality than existing options (like eBay relative to traditional collectible/antique dealers) can also grow the market.

While there are a few ways to size a market, we like seeing a bottoms-up analysis, which takes into account your target customer profile, their willingness to pay for your product or service, and how you will market and sell your product. By contrast, a top-down analysis calculates TAM based on market share and a total market size.

Why do we advocate for the bottom-up approach? Let’s say you’re selling toothbrushes to China. The top-down calculation would go something like this: If I can sell a $1 toothbrush every year to 40% of the people in China, my TAM is 1.36B people x $1/toothbrush x 40% = $540M/year. This analysis not only tends to overstate market size (why 40%?), it completely ignores the difficult (and expensive!) reality of getting your toothbrush into the hands of 540M toothbrush buyers: How would they learn about your product? Where do people buy toothbrushes? What are the alternatives? Meanwhile, the bottoms-up analysis would figure out TAM based on how many toothbrushes you’d sell each day/week/month/year through drugstores, grocery stores, corner mom-and-pop stores, and online stores.

This type of analysis forces you to think about the shape and skillsets of your sales and marketing teams — required to execute on addressing market opportunity — in a far more concrete way.

It is important not to “game” the TAM number when pitching investors. Yes, VCs seek to invest in big ideas. But many of the best internet companies sought to address what appeared to be modest TAMs in the beginning. Take eBay (collectibles and antiques) and Airbnb (rooms in other people’s places); in both these cases, the companies and their communities of users took the original functionality and dramatically expanded use cases, scaling well beyond original market size estimates.