The economic logic behind tech and talent acquisitions

Once upon a time I was a VP at a bank that no longer exists, but was prominent, and saw decisions being made the same way about whether or not to build a custom system or buy a vendor based system. Banks can gain a significant advantage over their competitors if they can build lots of great, custom stuff. Goldman has guys that build custom JVM's, for example.

The vendor based system promises to deliver some level of infrastructure and some level of completion for a project right out of the gate. It was fascinating to see the discussion happen.

The bank was considering two vendors. One wanted to sell their system for $10M and the street viewed it as having excellent quality. The other vendor was willing to sell for $5M and there was work that would have to be done at the bank to complete a lot of the project. The bank chose the cheaper vendor, thinking they could build logic to fill the gaps for less than $5M.

Scope creep eventually came in and the project turned into something that the bank spent well over $100M on. The project was not complete by the time the bank collapsed, but our new owners took the project on and spent more money on it.

$100M didn't seem to bother anyone as long as the people using the systems made a lot more than that. And to think, the decision was originally made based on whether or not they wanted to spend an extra $5M... It was said that the more expensive vendor was an order of magnitude better, so the squabble over $5M really seemed silly in retrospect.

Chris's point is similar to the lesson I learned at this bank. The lesson is that economics is a study based on relative notions, not closed systems. If you look at just the raw numbers, you might be coerced into bad decisions. I suspect we could've gotten better work done if we were on the better system.

I think this point can also be observed in the sale of Instagram to Facebook. Consider the balls on that team to ask for so much money! At first many folks said, "$1B!?" But if you break it down into a percentage of what Facebook was worth at the time (~1%), the numbers seems much easier to swallow.

1% of Facebook for the fastest growing competitor seems reasonable enough to me.

> Suppose you could build the product for $50M with a 50% chance of significant delays or failure. Then the upper bound of what you’d rationally pay to acquire would be $100M.

Am I the only one who thinks cdixon is slightly off here (except Chris with his reference to sophistication)? But what is the right way to calculate?

To simplify ignore the risk of delay for now (so 50% is risk of failure). Also assume the risk of acquisition is 0%. Then you have three outcomes: 1) acquisition, 2) internal success and 3) internal failure. In economic terms 1 means you gain $500M - X where X is the acquisition cost, 2 means that you gain $450M and 3 means that you lose $50M.

In this oversimplification I would say that the expected outcome of internal development is a gain of $200M (50/50 chance of $450M win or $50M loss). It would be rational to pay X < $300M since that would give you a certain gain of more than $200M.

Right? This is of course also lacking in sophistication but I have a feeling it touches on the angst that drives acquisition valuations up... :)

The interesting part comes in when people are looking outside their core competencies. So BigCorp's M&A department is trying to evaluate the purchase price of target A based on a complete lack of understanding on what sort of effort is involved in building Target A's product. When that happens two additional 'bad' outcomes are added to the mix, first is that the company passes on the acquisition for their internal project which they spend more money and more importantly more time on, or two they over pay for something that is easier to do than they estimate. This is where people like Mark Lynch can really sell the 'value' of something like Autonomy to a company like HP which has no clue how to build something like that on its own.

Few companies internalize the opportunity cost of doing it on their own. If you take your best performers and squirrel them away on this secret project to build Capability X, those same performers won't be out in your main engineering group catching small problems that will become big problems. So engineering gets a bit more chaotic.

All in all, it is a complex calculus that I have yet to see work out the way anyone thought it would.

Tangentially, I thought Clayton Christensen makes a compelling argument that this type of analysis—focusing on IRR and returns on equity—creates an overabundance of "efficiency innovations" that liberate capital but don't really push things forward. Perhaps the wave of acquisitions is tamping down our ability to produce longer-term, "empowering innovations." (http://www.nytimes.com/2012/11/04/business/a-capitalists-dil...)

The biggest problem now is, market capitalization is just a number on paper as it can't be converted to cash easily. So essentially they are spending cash to get more value on paper.

The only time market cap helps is when they sell their share holdings, or issue new shares. Because what they are doing is to convince people to pay more cash for a share, so that they can sell to them in someway.

Also the price-to-revenue ratio is not a magic number. It's most likely to be implied from the price and revenue rather than a valuation measure. It measures profit margin potential and growth potential. (Very) fundamentally the intrinsic value of a stock is made up of 1) risk-adjusted and growth-adjusted dividend payments and 2) cash-on-the-spot upon liquidation (so called "book value"). Price-to-revenue ratio is far less correlated with these two compared to price-to-earnings ratio or price-to-book ratio.

Acquiring a company: 1) may not help EPS (most likely hurt in the short term), 2) possibly help growth, and 3) may hurt book value. That's why for most acquisitions the acquirer's stock price is likely to fall by a small amount immediately after announcement. The exact market reaction will depend on the synergy of the acquisition, i.e. the strategic benefits in the long run (like reduced competition or cost savings).

I was hoping for a bit more from this article... didn't seem to be any big insights or interesting case studies, just a common sense explanation that didn't seem necessary.

Talent acquisitions, however, are not product acquisitions. The latter are easier to reason about in terms of business/economics, but the former seem a lot "squishier". "You hire an awesome team and then put them to work on something else" seems like it introduces too many variables to give you a lot of certainty about the outcome.

what is the "price-to-revenue ratio"?

When considering an 'acqui-hire' where the product itself gets abandoned, another factor to consider is the value of a 'gelled team'. (I'm surprised this isn't mentioned in the discussion at Dixon's original post.)

While in any acquisition there's a massive risk -- perhaps >50%? -- that the group won't thrive in the new corporate environment, they have at least, via their prior delivery of a competent product with some traction, shown that they can work together to create something real.

Picking the same N number of people off the street, even N great people, might not fit together productively with even a 50% chance.