Valuing High-Tech Companies

Author: Marc Goedhart, Tim Koller, and David Wessels (McKinsey & Company) · Published: February 2016, McKinsey on Finance Number 57 (Winter 2016) · URL: https://www.mckinsey.com/capabilities/strategy-and-corporate-finance/our-insights/valuing-high-tech-companies

One-line: Discounted cash flow still works for hot start-ups if you run it backward: size the future market first, then interpolate to today, and weight a handful of scenarios instead of pretending to a single number.

Summary

The article is an excerpt from the sixth edition (2015) of the authors’ book Valuation: Measuring and Managing the Value of Companies, published as investors were again “piling into shares of companies with fast growth and high uncertainty” and as the SEC announced in late 2015 that it would look into how mutual funds arrived at widely varying valuations of private tech companies. Goedhart and Koller were at McKinsey (Amsterdam and New York); Wessels taught finance at Wharton.

The argument is that the old tool still wins. Multiples such as price-to-earnings or value-to-sales are “of little use when earnings are negative and when there aren’t good benchmarks for sales multiples,” and they cannot capture what drives a particular company’s value. Discounted cash flow keeps working because “the core principles of economics and finance apply even in uncharted territories.” The pieces are the same as for a mature company, but the order changes:

• Start from the future. Picture the company once hypergrowth has settled into moderate, sustainable growth. Define that state with operating measures such as customer penetration, revenue per customer, sustainable margins and return on invested capital (ROIC), then decide how long hypergrowth lasts. For most start-ups, stable economics are “at least 10 to 15 years in the future.”
• Size the market, product by product. The worked example is Yelp, using public 2014–2015 data. Revenue grew from just under $26 million in 2009 to $378 million in 2014, and local advertising brought in $321 million of that. What matters for value is not unique visitors but turning local businesses into paying clients. Yelp’s target markets held about 66 million small and midsize businesses; 2 million had registered and only 84,000 paid. The authors model registration reaching 60 percent (8.5 million businesses by 2023), acknowledge that this is “extremely aggressive” for most start-ups, and justify it by network effects: advertisers want the site with the most traffic and consumers want the one with the most reviews. OpenTable’s 60-percent-plus share in San Francisco is their supporting data point. With paid conversion of 4 to 5 percent and revenue per client rising from about $3,800 to $5,070, they get local-ad revenue of $2.2 billion and total revenue of $2.4 billion in 2023. As a sanity check, that works out to roughly a 4 percent share of a projected $60 billion online local-ad market.
• Estimate margins, capital intensity and ROIC by analogy. Current start-up margins say nothing about the long run, so look at similar businesses. OpenTable’s management targeted mature margins of about 25 percent, and Google suggests that level is possible. Monster Worldwide is the warning: its margins were near 30 percent before 2010 and fell below 10 percent under competition. Internet businesses need little capital (Yelp’s invested capital was 24 percent of revenue), so ROIC gets so high that “it is no longer a useful measure.” The real capital sits in intangibles like brand and distribution.
• Work backward to current performance. Decide how quickly the company moves from today’s numbers to the target state: how long fixed costs dominate, and what scale it takes before revenue outgrows capital. Historical financials mislead here because accounting rules expense intangible investment. The Amazon example: by 2003 it had a $3.0 billion accumulated deficit, after expensing $742 million of marketing and $1.1 billion of technology development between 1999 and 2003. Marketing ran 10 percent of revenue in 1999, against about 2 percent of revenue for Best Buy’s advertising. The authors argue that gap was brand-building, which means reported ROIC overstates the returns available to a new entrant.
• Weight scenarios. A few probability-weighted scenarios make the key assumptions easier to see than real options or Monte Carlo simulation. For Yelp: a base case ($2.4 billion revenue, Google-like margins, $3.4 billion equity value), a high case (conversion doubles to 10 percent, about $4.6 billion revenue, $5.0 billion) and a low case (weak international expansion, under $1.2 billion revenue, 14 percent margins like Monster’s domestic business, $1.3 billion). Weighted 10/60/30, that gives $2.9 billion of equity, or about $39 a share. A footnote adds that Yelp traded between $40 and $50 in the first half of 2015, then fell after a cofounder left and the company guided to slower growth.

The closing point is about sensitivity. If the pessimistic scenario were ten percentage points more likely, Yelp’s value would fall by more than 10 percent. For a pure idea, the numbers are brutal: a start-up that must invest $50 million for a 5 percent chance at a $1.2 billion business is worth $10 million, and half a percentage point less probability cuts that value by more than half. So it is “no surprise” that start-up share prices are volatile, and understanding what drives value across scenarios matters more than a point estimate. For Yelp, market growth and share can be forecast within a reasonable range. The hard-to-predict value drivers are conversion to paid accounts and revenue per account.

Full text

Archived privately against link rot: ../attachments/valuing-high-tech-companies/valuing-high-tech-companies.md. The original article PDF is valuing-high-tech-companies.pdf, and its one exhibit (Yelp’s market penetration against US local-ad spending) is rendered at exhibit-page-3.png.

Connections

• Valuation: McKinsey’s case for discounted cash flow on companies with no earnings. It is the institutional version of Buffett’s point that the only question is the amount and timing of future cash.
• Valuing Young, Start-up and Growth Companies: Aswath Damodaran makes the same case against multiples-and-target-rate shortcuts and adds the parts McKinsey skips: survival probability, total beta and the value of separate equity claims.
• How Startup Valuation Works — Illustrated: the founder-side counterpart. Anna Vital describes the negotiation and comparables method that this article says tells you little about what drives value.
• Discounted Cash Flow: the method being defended, run backward from a steady state instead of forward from history.
• Total Addressable Market: the Yelp walk-through is a clean example of sizing a market bottom-up (businesses, registrations, conversion, revenue per client) and then checking the result against a top-down estimate of the ad market.
• Multiples: explicitly rejected as a primary method when earnings are negative and there are no good benchmarks.
• Yelp: the worked example.
• Amazon: the example of intangible investment that is expensed, which leaves reported ROIC misleading.
• McKinsey: publisher (McKinsey on Finance).