Trending
Zillow recently made headlines (and lost over a quarter of their market value) by announcing an end to their iBuying program. Not only are they ending their program, but they are sitting on a bunch of properties that are underwater. They will realize significant losses in a very strong real estate market. What gives?
It is quite ironic that a company that has as much data as Zillow, and provides Zestimate values for millions of homes, has found a way to lose significant value within a bull market in real estate. There is so much to learn from mistakes, whether it be our own or others. This is no exception. In this blog post I share three lessons investors can take and apply to their own decision-making process.
Information Does Not Guarantee Success
Zillow has a lot of information about the real estate market. One could argue, since they are an aggregator of data, that they have more information than anyone else. One would think this information advantage would give them a business advantage.
It has been said that information is power. Perhaps we could state it a little differently, “information can be very powerful if used appropriately.” There are two assumptions with information that investors need to be aware of.
- Information should only be considered if it is in line with your goals and values. A long-term investor will not benefit from all the noise of daily market movements, news headlines and forecasts. In fact, that information is more likely to be detrimental than helpful.
- Information needs to be consumed and used responsibly. Reading a headline and making a knee-jerk reaction is not being responsible. Consider the source of your information (it may be biased) and make sure you reflect on what that information suggests and whether there is contradictory information available.
Algorithms are Subject to Human Biases
People love algorithms because they are mathematical and take away all emotion from the output. However, algorithms are built by humans, and therefore may be subject to bias. In the case of Zillow, they adjusted their algorithm several times in order to bid high enough to win the property. In other words, their original algorithm, that ensured profitability, was losing out to higher bids. FOMO crept in. Rather than recognize they can’t make money in this heated market, they simply adjusted the algorithm to up their bids.
The algorithm got them to the place they are today. But that is because it was programmed by a human being and adjusted when the results weren’t desirable.
Investors will often adjust their goalposts as well. We have a plan, stick to the plan and when it doesn’t work as we hoped, we change the plan. Zillow’s initial algorithm may have been perfect, just like an investor’s original plan may be perfect. But since our evaluation time horizon is so short, we mess around with things way too often. We may be changing something that would have helped us be profitable in the long term simply because we lack patience and discipline, which often results in poor performance.
Consider the Error Term
Into each algorithm is built an error term. This term represents those things the algorithm cannot account for, such as surprise events. Error terms need to be sufficiently large to account for uncertainties, or the algorithm will produce undesirable results. With Zillow, their error term was too small (this is very common among Wall Street “experts” as well). A Zillow spokesperson said, “Our observed error rate has been far more volatile than we ever expected possible…”
The error rate for investors expresses itself in fluctuation. As humans, we love to project outcomes with certainty, or at least with a very small range of possible outcomes. But the world is uncertain, so our assumptions need to reflect that uncertainty – as uncomfortable as that might be.
Investors should expect a range of returns commensurate with their portfolio. While most investors may consider a drawdown of 10% acceptable, if it is a growth allocation, they need to also consider a drawdown of 35% as not just acceptable but highly probable to occur at some point. If they don’t, they are likely to “adjust their plan” (a nice way of saying “go to cash”) at the exact wrong time.
Concluding Remarks
Learning from others’ mistakes is a lot more fun than learning from our own. But it really doesn’t matter who made the mistake. What really matters is whether you learn from them. Learning is more than just recognizing a mistake and talking about it with others. Real learning is incorporating this into your plan, perhaps developing a defense system or decision tree, to ensure that you don’t make these mistakes at some future point.