You are using an older browser version. Please use a supported version for the best MSN experience.

VP Of Engineering At Quora Gives Three Vital Tips For Building A Strong Engineering Team

The Huffington Post The Huffington Post 2/03/2016 Quora
STARTUP © alvarez via Getty Images STARTUP

These questions originally appeared on Quora - the knowledge sharing network where compelling questions are answered by people with unique insights.Answers by Xavier Amatriain, VP Engineering at Quora, former Netflix recommendations, researcher, professor, on Quora.Q: What are the most important things for building an effective engineering team?

A: There are many things that an engineering organization or leadership team needs to take into account in order to build a great team. However, if I had to pick the 3 most important things I would go for the following (in no particular order):

(1) Enable and demand individual impact

There is no better way to have an "effective" team than to make sure that each and every one of its members has the tools and the space to have great global impact. There are many things that go into making that possible, including the two below: culture and diversity. However,  in my opinion, the most important foundation to have engineers have the right amount of impact is to have great engineers to start with. That's why hiring plays such an important role (if not the most) in building a great team. If that piece fails, there is not much more you can do to fix things.

I am a big fan of "small" engineering teams precisely because they enable engineers to have the right amount of impact. I am always suspicious of people that brag about "the size" of their team or fix things by just throwing more engineers at the problem. Both Netflix and Quora are very lean organizations where engineers have a huge amount of impact, but I know that is not the case for most other companies.

(2) Work on culture

I am also a firm believer of having a clear company/engineering culture that makes teams feel aligned and ensures everyone is not only working in the same direction, but also has a shared understanding of how "things work". Over time I have come to realize that it is not so much about "the particular" culture you have, but rather about the clarity you have and how strongly you motivate it and even enforce it in your day to day work. If you look at successful companies like, say, Google, Apple, Facebook, or Netflix, they all have extremely different cultures. However, they all are similar in the fact that they are very clear and even passionate about their culture. Having a clearly defined set of values and cultural environment enables better communication and makes engineers more comfortable with taking risks and anticipating the result of their actions. Some people are now calling this "psychological safety" although I think it is a quite misleading term.

That said, different cultures will attract different kinds of people and not all cultures might work depending on your business,  strategy, or company situation (e.g. startup vs. large multinational). So, yes, you should pay attention to those details. But, above all, you should be clear on your culture and values and make sure these impact everything from hiring to compensation.

One particular cultural aspect that I think is very important is to avoid having "brilliant jerks" in your teams. As much as you want to enable individual impact and hire very smart people, it is very important that those individuals can also enable others to have impact and be successful. Brilliant jerks might become successful themselves, but they will become a cancer in your organization and will make everyone else fail. Avoid at all costs.

(3) Ensure diversity

One of the risks of hiring very smart people that are all perfectly aligned to "your culture" is that you end up having a team of clones that all look very much like each other. It is well known though that diversity improves performance of teams. I do think that ensuring diversity is one of the most important yet challenging  parts of creating a great engineering team. Since I am a big soccer fan, I always use the metaphor of a soccer team to explain this (I realize sports teams are segregated by gender and in that sense are not a great example of diversity, but please bear with me on that aspect): if you are building a great soccer team, you couldn't do it by having 11 Messi's although he is the best player in the world (see My answer to Who is the all time best footballer?  ). You need players that can score goals, but also good defense, a good goalie... This holds even more so when you are embarking on a creative endeavor such as engineering. The more diversity you cultivate, the easier it will be for anyone you add to be welcomed and valued and the better off your whole organization will be. I am glad that as challenging as this is, we are talking more about it and there are now better tools and resources available (see, for example, this document by Homebrew).

Again, many other details are important, but the three outlined above should set a great foundation to get started and will even drive other aspects in the right direction.


...
Q: How do you deal with strong difference of opinion among expert engineers in the team?

A: This is a great question. However, the question details contain an assumption that I disagree with: the way to deal with strong differences between engineers is not  necessarily to build consensus but to have a clear owner in charge of making the final decision.

First, let me say that it is great for an organization when people not only have opinions but feel empowered to express them, even strongly. Conflict avoidance is in fact one of the leading causes of poor decisions and ultimately organizations and teams failing. It is good to encourage discussion especially when a technical decision needs to be made.

However, I have also seen situations where two or more engineers came in with radically different ideas and there was no way to come to an agreement. I have even witnessed cases where this kind of situation ended up leading to pretty bad outcomes. In all of those cases, looking back, the problem was that the decision process, including the owner, was not clear.

Here is the way to address technical decisions/discussions (or most decisions, for that matter): it is called the consultative model. Whenever a (complex) decision needs to be made, start by defining who will be the owner. The owner will ultimately be in charge of making the decision. That person will also be made accountable for the results of the decision. In order to make the best decision possible, this owner will be encouraged to consult with as many engineers as they see fit. Engineers involved in the discussion will be encouraged to express their opinion as strongly as they want. However, they will all accept that it will be the owner the one who will make the final decision. It is up to this owner to make sure the different parts feel like their opinions are being taken into account. However, the goal is not build consensus. The owner can, for example, say things like "I totally get what you are saying, and this is a valuable opinion. However, we are going to go for the other option because of X and Y".

A couple more things are important when you implement a model like this one:

First, you need to make sure people understand the meaning of "decide and commit". This means that once a decision has been made, regardless of what your opinion was during the discussion, you should commit to what was decided and work with the group to implement it. This is a really key trait of great team players, and one that can be learned, by the way.

Finally, in this model you should also account for bad decisions being made. As I mentioned before, the main correction mechanism is holding owners accountable for their decisions. However, sometimes it is best not to wait for the outcome of the decision because it might be very costly. It is important that just as engineers feel empowered to voice their opinions during the discussion phase, they should also feel empowered to escalate concerns whenever they worry that a bad decision was made. Escalation is not a failure of this kind of system, it is a feature that should be used as much as needed.


...
Q: Should we worry about the "filter bubble" created by recommender systems?

A: No. Really. You have more important things to worry about. Three reasons why you should not worry about the filter bubble:

Editors are much worse than algorithms

When I was growing up in Spain we had only two TV channels, both of them controlled by the government. That is a "filter bubble", not what recommender systems create on your homepage. Even nowadays, most newspapers have an "editorial line" that is far more skewed than any algorithm. That's actually why people that need or want to have a balanced view will read both the New York Times and the Washington Post (and as many other newspapers as they can). Compared to that, algorithms offer us a much more diverse view of the world.

We are much worse than algorithms

Left to our own devices, we are much worse than automatic systems or algorithms. I admit it, I will personally unfollow anyone on Twitter who supports Trump. The people that I have manually selected on Twitter are much more skewed than anything an algorithm can suggest. The same goes for the online newspapers that I personally decide to read. Algorithms, if anything, introduce diverstity and some level of exploration on my daily life. It seems I am not alone on these choices. See Computing Political Preference among Twitter Followers, for example.

There are worse things to worry about in recommender systems

Recommender systems can be subject to all kinds of biases. For example, a recommender system might be biased to recommend content that is economically more interesting for the system. Or, it might be biased to recommend items that are globally popular.

Just to be clear, diversity and exploration are issues in recommender systems. However, there are well-known approaches to both and it is usually in the interest of the algorithm to increase them in order to improve long-term retention and satisfaction.

And, again, there are probably much more important things you could be worrying about regardless of whether someone decided to write a somewhat popular book or give a TED talk on them.

These questionsoriginally appeared on Quora. - the knowledge sharing network where compelling questions are answered by people with unique insights. You can follow Quora on Twitter, Facebook, and Google+. More questions:​

More from Huffington Post

The Huffington Post
The Huffington Post
image beaconimage beaconimage beacon