Hiring a VP of Engineering or CTO For Non-Techie (First Time) Founders
So you start a web service with some friends. They do the programming. You do everything else. Suddenly the service takes off. You are thrilled and getting ready to pop open the bubbly. But then things start to sputter as your service can’t handle the load. You put the bubbly back in the fridge (unopened) and wonder what to do. The next day you hire a kick-ass VP of Engineering, the service starts humming again and everyone lives happily ever after. The End.
Except for the part in which you have no idea where to find a kick-ass VP of Engineering, or how to tell one apart from the hundreds of impostors at your door. So what to do in real life? Here are some do’s and don’ts (mostly learned the hard way) presented as (only slightly) exaggerated scenarios.
1. I just met this guy. He’s totally awesome. He can solve all our problems. I am going to make an offer right away.
Don’t. First, ask yourself, how many people have I met for this position? If your answer is less than five, stop right there. You need to see at least five candidates. And that’s an absolute minimum. If you are a first-time founder you most likely have no point of comparison and it’s too easy for tech candidates to look great when in fact they are not. Also, make sure to get as many experienced people involved in the process as you can. This is where you need to lean on investors, board members, advisors, fellow entrepreneurs, etc.
2. This woman is amazing. She spent the last 20 years at Monsterously Big Corporation in IT. She has soo much experience. I will hire her immediately.
Don’t. Most people who have been in IT at large corporations have either forgotten or never learned how to get things done on a budget of less than a gazillion dollars and a three year development cycle. Similarly, you should be wary of candidates coming from academia or a company’s research division. They tend to be too theoretical for building robust production systems. Look instead for candidates who have actually delivered at early stage, high growth companies, ideally already in a VP of Engineering / CTO role.
3. Met this candidate. Blew me away. Our approach is all wrong. We have to rewrite everything in [insert programming language / framework here]. That’s how they do it at [insert successful startup here].
Don’t. Your candidate is likely to be a technology fanatic. Advocating a complete rewrite without having seen the existing code is not a sign of pragmatism, which is a key trait for actually getting stuff done. In any case, almost anything can (and has been) built in any language and complete rewrites that never really complete have been the kiss of death for many companies.
4. This engineer was a major contributor to a [insert open source project here]. He wrote a really cool mash-up using our API. He has great experience in a high traffic environment. Should I hire him even though we don’t yet have the VP of Engineering?
Do. Great engineers will be sufficiently self-directed. Also, there is likely to be much to do if you are really growing, from making small changes to improve performance (those always exist, by the way) to adding new features or writing ancillary systems. Even if you do find a great engineer, you should, however, resist the temptation to make them the VP of Engineering or CTO. It takes a pretty different skill set to manage people and process than it does to build stuff. That is not to say that a great engineering hire can’t step up and fill that position, but it’s much better to let that happen than to force it (in the latter case if it doesn’t work, you have just lost a bunch of time finding the right person and most likely lost a great engineer in the process).
5. We have seen tons of candidates. It was an exhaustive and exhausting process. We all love [insert name of kick-ass VP of engineering candidate] but she has three different companies wooing her. I want to make an aggressive offer with more equity than we had originally planned for this position.
Do. This is a clear case where the right person will grow the pie by a lot for everyone (in fact, there may not be a pie without the right person), so giving a bit more away makes total sense. Also, you will still want to have a 1 year cliff for the equity vesting, in case things don’t turn out as hoped for after all. If you really have to, in order to land the candidate, you might go down to a 6-month cliff, since you will (or should) know after a couple of months whether you made the right hire.
P.S. If you are a VP of Engineering or CTO with the right stuff, please write to us at info@unionsquareventures.com – we love to develop relationships for when the right opportunity comes along.
August 7, 2007 08:43 PM, By Albert Wenger
Tags: cto engineering howto startups vp
Comments (24)
I'm biased on this because I've been an engineering manager for a while (so I'm probably recommending things that fit more with my own skills), but still I'd recommend the following few tests if you're non-technical and looking for a technical leader:
(1) Recidivism rate: there are a lot of engineering managers who have worked with particular engineers at several different companies. Ask if there are any people they would want to bring along should they be hired, and why. Having a loyal team across companies suggests that they develop people's careers rather than burning people out to meet a short-term need.
(2) UN translation skills: take *any* technology in your company and ask them, off the top of their head, to explain it very clearly to your least technical executive. Repeat that test several times during the interview process. Great engineering leaders can communicate technical issues and needs throughout the company.
(3) The magic triangle: Talk to them about tensions between product quality, cost, and schedule. If they swear on a stack of bibles that they never miss a deadline, to use Albert's phrase, "don't" -- at least not for a new startup. New product development usually involves enough uncertainty, especially if the technical aspect of the company is complex enough to demand a great engineering lead, that precise estimation is impossible. That said, you should be comfortable with the tradeoffs the engineering lead will recommend, so if their discussion of these tradeoffs gives you the willies, you may need a different person.
Those are a few good ways to get started. If at all possible, put the candidate in front of other engineering leaders you know and respect -- there's a lot that a technical person might catch, so having a friend who can confirm your gut feel is important.
Good luck!
Posted by Marc Hedlund , August 8, 2007 12:15 AM
Great post. I think there is a related post opportunity for techie types that need to hire a CEO or business/sales/marketing VP.
I'm in the latter group of people and frequently meet non-technical folks who think that tech is just plug and play ("I don't want to pay/give up/invest much because it should be easy").
Another suggestion to consider: sit down with a couple of development firms and see what _they_ think it will take to develop something and why. It can really help a non-techie wrap their head around the variety of challenges and complexities they haven't considered. If you don't want to take an in person meeting, put your project on one of the freelance coding websites (like http://www.scriptlance.com/) and see what feedback you get.
In my current startup (not my first by a long shot) I got very lucky and met a great CTO candidate just as his company was offering a severance package. Nothing like having a CTO with an obligation free paycheck for 3 months! We've been rolling along nicely for 9 months now.
Posted by fewquid , August 8, 2007 07:57 AM
Great post, although I would argue you can take most of the points and apply them to a VP of Marketing hire or a VP of Sales hire as well.
Another big qualification that I would look for is has this person every been responsible for developing and _shipping_ a commercial product.
There is a HUGE difference between writing an internal software application in a typical IT environment versus having to ship a product with real paying customers. In a commercial development shops there are marketing pressures related to both schedule and features based on how the marketing team is positioning your product. There are competitive pressures as you look to gain or maintain a first mover advantage. And, especially in the startup world, there are serious financial/cash-flow pressures that mean you need to ship your product on time.
So when looking for a VP of Engineering/CTO, make sure it is someone who has had to deal with these factors. Most people would be surprised to learn how few technology people have ever really had to ship a product under the duress they will have at your startup. That even includes senior managers at what you would consider to be technology companies. If the candidate has never been responsible for really getting a product out the door, they are probably not a good fit for a VP role.
One other comment. In the situation described above (startup without a serious tech guru is getting traction and needs help), I can make a couple of assumptions about the technology environment at the organization:
1) The current technology is/will have serious scaling and maintainability problems due to a lack of initial architecture planning. You probably know this already.
2) Rewriting everything from scratch is probably a bad idea. For a much better explanation, read Joel Spolsky's article, Things You Should Never Do (http://www.joelonsoftware.com/articles/fog0000000069.html) or Chad Fowler's The Big Rewrite (http://chadfowler.com/2006/12/27/the-big-rewrite) for a much better explanation of why..
One of the major challenges that your VP of Engineering will likely have to undertake is dealing with these two points by coming up with a plan where you will be able to lay in a new and improved architecture while still working around the current system. Generally, this will mean developing a strategy to get the better architecture in place on a module by module basis, making gradual improvements rather than going for a "Big Bang" development effort. When interviewing prospective candidates, I would try to get them to discuss what kind of approach they would take to solving your problems. You will be surprised how quickly a good VP candidate will grasp where your problems are (both in the technology and in the people) and can start analyzing how they can potentially be solved.
Posted by Rob Di Marco , August 8, 2007 08:49 AM
No really competent CTO is going to work for you. This startup is too far along for them to get a founders stock package. Those friends doing the programming probably have more stock than the new CTO will get. Unless it is pretty obvious you're on a rocket ride to the moon and they want to hang on; they'll clone you and start their own version.
Posted by Jon Smirl , August 8, 2007 09:55 AM
Great points, Rob. One of the other intangibles to look for is process. Like architecture, it's important to understand the full range of "no process" to "mature process" and be able to apply the right principles at the right time without interrupting development schedules.
As a follow-up to "have you shipped a commercial software release," you can ask about things like the following:
* How did you know it was time to ship? Look for an understanding of the bug tracking process, quality management, and business tradeoffs between quality and time-to-market. Key points: bug tracking, automated testing (unit and functional), quality gates.
* How did you handle service packs / fixes? Look for an understanding of the customer service process, configuration management capabilities, and multitasking. Key points: source code control and software build capabilities that support branched-code maintenance and delivery.
* How did you know what features to leave out? Look for an understanding of product management, project management, and customer value. The VP of Engineering won't necessarily be the point person for these decisions, but he needs to be able to participate constructively in them.
* How did the product make your customers (or Professional Services group) more productive? Again, this is a softer point, but you want to see an understanding that the software isn't worth a hill of beans until it makes someone happy. How did the software make customers happy?
Posted by D. Lambert , August 8, 2007 10:42 AM
The options are very much dependent on funding. The original description sounds like a bootstrap deal, but scenario 1. mentions leaning on investors. If you have funding from a VC your task will be easier. If you are bootstrapping, I would go for a director level candidate that has managed a group of up to 10.
Do not consider someone that wants to work in a start up but has been at a big company for all of their career.
The best hire for you will be one that is within 3 degrees in your network.
Most important qualification: what product have they shipped, did it ship on time.
reference: ceo they worked for, best Jr. eng they hired, best employee that was in the group.
More important references are the BACK DOOR. find people in your network that you can do an un-official check. When you find the candidate that passes all of the above, move quickly and make an attractive offer. This hire will determine the quality of all future hires you make in your engineering team. Start with good DNA
Posted by REM , August 8, 2007 10:51 AM
Consider this one of the most important hires you will ever make. A bad move might very well kill the company.
The VP Engineering will build out the team going forward (also Ops, Test, etc) and most likely pull in folks they have worked with in the past. Find out the last time they actually wrote code that went into PRODUCTION. They might not be coding every day as the VP Eng, but in a small startup, they need to be able to roll up their sleeves.
Posted by Dave Hodson , August 8, 2007 10:59 AM
I agree with the rules above for avoiding bad situations, but there's not much discussion or advice on what to do in the happy situation where you actually find a good candidate. What does it take to attract that candidate to your company?
The setup in the first paragraph is perfect, because most of the time the req for a VP Eng gets opened up only when there's a problem: service isn't scaling, 2 lead engineers are at each other's throats, VP Sales is screaming bloody murder that product isn't shipping on time, BoD has lost faith in the last guy...
My advice: when talking to your dream candidate, don't lead with all the problems, lead with the opportunity, and most of all with the product. Financial opportunity is nice, but the reason a VP Eng gets out of bed in the morning to to ship good, working, valuable product to happy, delighted, overjoyed (and, ideally, paying) customers.
If the job description is nothing more than "fix all these nasty problems until the service hums and we all get rich", you'll probably lose. There needs to be a vision towards which the product/service is striving, and the VP Eng needs to be challenged to realize and expand that vision through execution and innovation.
The innovate/expand part is key - it's the fun stuff you get to do once the basic execution is under control. Fixing all the problems is interesting and challenging as well - in spite of all of of them, being a VP Eng really is the greatest job on the planet - but there needs to be an upside of fun, exploration and innovation to make it all worthwhile.
Posted by Alan Steele , August 8, 2007 11:46 AM
I ardently hope I never have to recruit another VP Engineering, ours is so wonderful (and I think happy). But I'll bookmark this discussion for other less-fortunate friends.
Posted by Robin Wolaner , August 8, 2007 01:13 PM
Three comments:
(0) It is worth belaboring distinctions between a CTO and a VP of Engineering. Many people will see the CTO as a technologist, and the VP of Engineering as a manager. These are different people, almost different species, and if you need one person to wear both hats, you may need an inflection in one or another, and this can be a big, big deal. I generally agree with Albert that hiring the VP from GinormousCo is almost surely always a mistake, but it is possible to have an early stage startup where the crucial issue is about technology management: E.g., there are technology integration points that are more about people and processes that the technology per se.
(2) If one of the founders is not a CTO or a VP of Engineering type for a web-based startup, good luck to you. The chances of being hoodwinked, swindled, abandoned, lied to, abused . . . You really need a spolid person in that role right away. If you don't have one, get your VCs to find someone who can do it for 3 to 6 months to get you to the point where you can "sell" to a real CTO or VPE. 'Cos you're just going to be a babe in the woods. I don't care if your MBA came from Sloan, you're just not going to have the means to figure out who the good CTO/VPE is, and if you luck out, you'll still be worried. Frankly, I'd be worried about a VC firm that would invest in a web-based startup that doesn't already have a CTO/VP of Eng. type in hand.
(3) In the early stage (fewer than 20 people in the whole company), it can really pay to have a CTO/VP of Engineering who can literally do it all: Provision equipment, design and plug together the equipment (or find the right hosting service, etc.), help with requirements, communicate with potential customers in everyday language, write code, figure out bottlenecks. Maybe not brilliantly, but they have to know the whole stack. Why? Because on a small team (say, 5 in engineering), it can happen that 2 will leave, and 1 will be out for medical, and you're on a skeleton crew, and someone has to answer the questions. There are such people out there. The hard part is that this person has to understand the dynamic of replacing these facets of his job with specialists as the team grows.
In my experience, these "do it all" types can be young, they can be old, they may have a core in traditional "IT," they may come from the ranks of the developers: It's very unpredictable. But they are worth gold.
Posted by John , August 8, 2007 05:39 PM
You just described the first seven years of my life trying to put together a great technology team for my company. It finally works very well and we are pretty dangerous. That post was awesome and we are all chuckling in the office as we travel down memory lane.
Posted by Michael , August 9, 2007 10:41 AM
I must say that this is a fantastic post. It isn't often I come across such an informative post and comments. This is a great post for things to think about as a person who is looking to launch a company and how important having a technical co-founder is. Or alternatively, an established company leader in avoiding making a bad technology leadership hire.
Posted by Lou Paglia , August 9, 2007 04:21 PM
The VP-Eng & CTO are 2 different job types.
VP-Eng Responsibility: To deliver products.
He Speaks: Budget,resource,process,schedule and product delivery.
CTO Responsibility: Technical roadmap for the company.
He Speaks: Evangelism, Stds, Architecture, Technology, Vision,Customer...
Typically tech founder will hold both positions for phase-1 and latter company will hire a VP-Eng.
It's hard to get a person with both qualifications and if you found one, he might already be developing his own business plan.
I agree with one of the previous comment here:
No really competent CTO is going to work for you....they'll clone you and start their own version......
So, you may want to hire a VP-Eng first!
Posted by Uday Subbarayan , August 10, 2007 07:34 PM
Tech teams are always tough to bring together. At my old company, I knew someone that was in your shoes and let me say, the stress defintely showed on his face! I'm wishing you the best of luck :)
Posted by Tom , August 26, 2007 08:16 PM
Excellent blog - I can relate with so many points identified by all contributors. Having been in the role of technology leadership, I agree that a similar blog would be nice for a focus on VP Marketing, Sales, and/or CEO.
I currently work with a startup that was abandoned by the original Technology leader only months before the service was scheduled for release. I stepped in and gained control of a platform that was completely undocumented and designed in a vacuum (the departed partner was in sole control of the technology). The other two partners were left in a lurch, and if someone such as myself had not stepped in, the business would not have launched.
So, in addition to all the other good advice within this blog, I highly recommend taking the time to document the technology as much as possible. You never know if someone new to the team is going to have to assume ownership, and you will be doing him or her a big favor.
Posted by Tad Rodgers , August 27, 2007 12:15 PM
I think you are also missing that part about the fact that the best tech people are not the best salesman. If someone looks really attractive (in a professional sense) than they are probably great at career advancement and sales.
Instead I would look for referrals, and not just any kind of referrals, but referrals where the person is known through a very close (family friend, father, etc...) relationship. Why? Because these tend to be sincere.
Posted by China Internet Marketing Blog , September 6, 2007 09:04 PM
Excellent Article opened interesting discussion among the community. Focusing on the topic of discussion.
Here are my 2 cents.
VP of Engineering,
1. One who wants to be new VP of Engineering, after working for several years as team lead or a senior engineer etc..
2. One already is a VP of Engineering, and wants to move to same position.
the first category of people (once me) are aggresive in taking up the workload. They are handson and try to impress the founders agreeing to the all they want to see. Often loosing one team member creates a choas until a new person has been found to replace, spear the load and cut the doc .. process. So, documentation etc. are on the backburner (we will get it done when we have more time !!!, that generally never happen until 2 years into the project)
the second category, are bit experienced in the game, less aggresive in taking up work (burned once or twice before). Would want to be on the conservative side. They insit on documentation (they prefer to takeup less tasks/scale down the features) They would join only when they see "gold" in your company. If you have hard time selling your idea to a "VP" of #2 case, go for category #1. Back him up with "specialists" on consultant list.
Have a good recruiter inplace. It is work paying the money to a recruiter for a key position like VP of engineering, don't count too much on your people skills or your job postings and screening skills.
Posted by Ramesh Kasavaraju , September 11, 2007 06:58 PM
http://imerl.fing.edu.uy/calcu1/_discusion/_disc1/0000120c.htm?female_viagra
Posted by Nina , November 18, 2007 04:41 PM
http://pills24.meds-store.org/rendezvous-allegra.html
Posted by Zinaida , November 22, 2007 08:58 PM
http://pills31.meds-store.org/cipro-support-forum.html
Posted by Olesya , November 23, 2007 10:15 PM
Great post! I've worked in startups for the last 7 years and have already seen most of the cases you discuss here.
I wrote a small post on the subject a while ago focusing more on the people-management approach rather than technology issues and brilliancy of the CTO. ( http://thebigdeal.wordpress.com/2007/11/26/celebrate-diversity/ )
A good CTO/VP-engineering has to be able to handle a bunch of brilliant developers, who, as such, are certainly going to be self-absorbed egocentric people; meeting somebody who can keep a team together and each single individual productive and challenged is surprisingly difficult.
One thing we used to do in every company I worked for was, after an initial selection of the top candidates amongst the mob, having every member of the existing team interview them. Not so much to test the candidate's competence but more to check whether we all would have been comfortable working with them. Something often underestimated but incredibly important, especially when your team is just over 10 people.
Posted by Stefano Buliani , January 9, 2008 09:39 AM
Albert:
I've linked to this post at my blog for the Innovators-Network in hopes that my readers will visit Union Square Ventures and read your complete thoughts on how best to find a VP of Engineering or CTO for a growing web startup.
Thanks for your informative and insightful post!
Best wishes for the New Year!
Anthony Kuhn
Posted by Anthony Kuhn , January 9, 2008 06:02 PM
Wow, cool man, big thanks! http://heztlpftxc.com
Posted by prkyjezgkv , May 8, 2008 05:18 PM
At first, I thought the blog was going to give some technical advice on what to do when a web service takes off and everything is crashing because it can't handle the load. I am taking a mobile security startup through this exact scenario - moving from a dinky shared Windows server use for .Net app development, up to a dedicated Windows server, simultaneously developing the GlassFish/Java image for use on EC2. Amazon EC2 is the way to go if you need to scale BIG, FAST. I am already using their S3 storage, but the move to EC2 is more complicated.
But of course, the blog was actually about hiring someone to do all this. Good luck!
Posted by s , August 7, 2007 11:18 PM