Tips for working with offshore teams

Remember “The World Is Flat”? For software development, the world is definitely getting more planar. I’ve had the privilege of working with a few offshore software development teams directly, and have discussed the practice with other engineers.  I wanted to share some of my tips. My experience has been in web development, but I believe these tips apply to most general software development. (If you’re looking for a tiger team/extreme expertise in a specific area, I’d treat an offshore team just like another vendor.)

If you can, pick a team that you have worked with before. Just like onsite software development, knowing personalities, strengths and weaknesses of team members is crucial to delivering quality software. If you haven’t worked with someone before, then ask for referrals and check references.

Pick an easy first project. This could be something not on the critical path, a rebuild of outside functionality, or a rewrite of an existing piece of software. If you have a prototype that you can point to and say “make it work like that”, that can reduce complicated requirements discussions. One successful offshore project I know of took a Cordova mobile application and rewrote the app in objective c and android java.

In general, projects that have a lot of iteration and experimentation are tough to offshore, because of communication latency.  You lose out on feedback cycles.

Have a working agreement. This can be an informal document, but you want to specify roles and responsibilities (who is releasing? who signs off on stories that have been finished? how does planning happen?). Make this document a living one.

Having regular time overlap can speed up feedback, to some extent. If you are offshoring to a country in the same timezone, this is painless.  If you are working with a team from an offset timezone, you may need to adjust your sleep schedule.  It is best if both parties have to adjust their lifestyles somewhat–it is more equitable. Record these overlap hours in your working agreement. Shifting my schedule to have a significant overlap doubled the number of feedback cycles (code reviews, questions) that can be made between myself and an offshore developer.

You need one person to own the relationship on each shore. If there are process questions, these folks sort it out. They can button up any lingering requirements uncertainties. They may pull other folks in to make decisions, but this pair owns the success of the offshoring relationship.

Leverage asynchronous tools. Make sure you use a progress tracker like pivotal or trello. slack is fantastic. So are the modern source code management SaaS applications like github and bitbucket. A live prototyping tool like Invision is useful. Record decisions in writing.

Finally, decide how closely you need to follow the code delivery. You may or may not want to see “how the sausage is made” This can be based on a number of factors:

* how much technical expertise does the onshore team have? Are they interested in acquiring more?
* how much time does the onshore team have? You don’t want the onshore team to be a roadblock.
* will the project be standalone (a marketing website)? Or integrated into the main codebase?
* is what the offshore team writing core functionality to your application and/or business?
* who is responsible for maintenance and changes after delivery?

Many of these tips are best practices that should be followed wherever your development team is.  But they are especially important when they aren’t in the next office.

Offshore development can extend your budget and pull in timelines.  It can help you build a better product, or allow you to access skillsets you might not be able to hire.  You just want to make sure it works.


How we work around Stripe’s seven day wait for new account payouts

Many of the kitchens on The Food Corridor are setting up their Stripe accounts for the first time.  They are coming to us to help with their billing, invoicing and scheduling issues.  One of the selling points is that we collect payment information from their clients and bill the clients automatically, helping kitchen cash flow.

However.

Stripe has the concept of a payment (which is when the buyer gets charged and money moved into Stripe’s system) and a payout (which is when the money moves from Stripe’s system into a seller’s bank account).  These are definitely not synchronous, for what I assume are concerns about fraud and money to be made from float.  The first time you charge against a Stripe account, they hold the funds for 7-10 days at Stripe before paying out.  As a user, you can see the funds, but you can’t access them.  (You can’t pay rent with a Stripe account balance.)  This was frustrating to many of our clients, and a horrible first experience with our billing system: “we said we’d help with your cash flow, and we will, it will just take a month”.

To alleviate this, as soon as their account is set up, we make a small charge against our own credit card and send it to them (a pre-charge).  This starts the clock on the 7-10 days mentioned above.  This happens before the first real client billing, which means that the first real client billing will be paid out in 2 business days.

One of the things I love about working on a product for a period of years is that you get to make these types of refinements which are not technically difficult, but truly matter for the user experience.


Distribution matters

No duh, right?  But if you are building a company, you should think about distribution in the same breath as you think about your product.  Because it’s part and parcel of the business.  If you can’t deliver the product, at a profit, you have a project, not a company.

From this blog post, which talks about building boring foundations instead of sexy new highly profitable features that have to go through gatekeepers.

But these payment processors had the customer relationships, and they had the starting product that the customer wanted.

See also this post, where Seth Levine talks about features vs products vs companies.  It’s fine to start a company as a feature, as long as you have a viable distribution channel.  But it’s more likely that customers will want products (though that depends on your customer–Twilio and AWS both have made plenty of money selling features to be incorporated into others’ products).


Run through the finish line

I used to run cross country way back when. One of the differences between the good runners (like me) and the great runners (Dave F, Todd E) was talent. But drive was more important than talent. One of the key pieces of advice my coach gave us was to “run through the finish line”. On the face of it, it seems obvious–you are at the end of a 5k race, why wouldn’t you want to finish as fast as possible? Why waste that effort? But I can’t tell you how many times I passed (and was passed) in the last 50 feet of a race. I saw the finish line, I relaxed (as much as possible) and slacked. Or, I saw someone else doing that and knew I could move up one more place. If you’ve left it all on the course (easier to do in a 5k than some other races) it is hard to have anything left to run that last bit. But drive can provide that last bit of energy.

In a software project of any size, there’s also that “last 50 feet”. If you are developing, it’s the last 10% of the project that takes 50% of the time. If you are running a project, it’s the last few weeks when all the little things that were put off during the big build out phase come out of the woodwork and need to be dealt with. Or it’s when the project blows up 2 weeks before deadline. When you are dealing with a vendor, it’s the crunch time before you launch, when any organizational complexity ignored comes due.

If you want to get across the finish line successfully, sound of body and mind, and with your team intact, what can you do to finish strong?

Be realistic: I never ran a 15 minute 5k. So starting out at that pace was a sure way to make sure I had nothing left at the 2 mile marker, let alone the finish. Make sure you have realistic expectations at the start of your project. If you or your team don’t know how to scope the work, spend some time learning how to do so upfront.

Cultivate drive: remind the team regularly why this project is important. If there are features that you discover are harder or less useful than supposed at the beginning, don’t be afraid to shift effort. Also, refer back to previous successes or failures for motivation.

Confront complexity early: don’t save it for the finish. That’s like picking up a 40lb sandbag at the 4k mark. Derisk a project by doing the hardest parts first. In some situations you can’t (final integration of components or external system access), but try to do it as early as possible (via CI or mocking up external systems and loudly asking for access as soon as possible).

Hold a reserve: experienced runners know to save a bit for the end. Don’t ask your team to work crazy hours in the beginning or middle of a project, because they’ll have no reserve to push through the finish line. Every software developer I’ve ever talked to knows the last 2 weeks of a project will require extra effort–don’t waste that expectation by requiring more work early on.

Drive is important to finishing a 5k; it’s also important to complete a software project. Running all the way through the finish line can be the difference between a failed delivery and a successful project. Plan accordingly.


Turning Zendesk tickets into help center articles

We used Zendesk as part of our ticketing system at The Food Corridor. (Other parts are phone calls and an email inbox.)

I’ve been pretty underwhelmed with Zendesk as a ticketing system–I find it hard to understand and the UX is overly complicated for our needs. However, we recently moved a large number of FAQs into the Zendesk help center, and I found that aspect of the software to be awesome. I’m especially excited to see if it cuts down on customer interaction by being integrated into the web widget.

There’s also a cool way to take comments on tickets (typically with answers to common questions) and turn them into help center articles. Here’s how you do that:

An admin needs to install this Zendesk application and I recommend setting the ‘draft only’ option.

Then, the steps are:

  • Go to a ticket.
  • Click on the ‘apps’ button in the upper right hand corner
  • Click ‘post article’
  • Select the comment you want to post (you can only pick one)
  • Enter your title
  • Modify html if you’re comfortable, but realize it will be a draft article regardless, and you’ll have a chance to change it later)
  • Click ‘next’
  • Choose a section of your help center
  • Click ‘post to section’
  • Click the ‘view’ button (for some reason the edit button didn’t work for me)
  • Click ‘edit article’ (upper left)
  • Modify the content if needed using the guide editor, which is nice and WYSIWYG
  • Change from ‘draft’ to ‘published’ to make the answer available

I remember a knowledge base project over 15 years ago that was aiming at sharing knowledge across our organization. It flailed. Zendesk help center seems to be well on the way to achieving such aims.


Ditching Hourly Rates

Back when I was consulting full time, I typically charged by the hour.  For some clients I’d do fixed bid pricing, based on an hourly guess, but that was typically after we had an established relationship.  Otherwise the risk of losing your shirt is just too high.  I’ve done that once or twice–no fun to be working to finish up a project and just knowing your hourly rate is heading far too rapidly towards single digits.

I’m not consulting now, but if I were I’d be following Jonathan Stark’s advice on value pricing.  You can sign up for his free email course, which is valuable.  After that you’re put on his generic email list which is 25% pitches for his business coaching and 75% good tips about value based pricing.

I will be honest, I’d have a lot of fear about moving to value based pricing, the same way I have fear about niching and focusing on a particular market.  Both are scary concepts because when I am a consultant, the feast or famine nature of the business makes me want to say yes to everyone (within my available skillset and time).

But these will be my first two business experiments if I ever go back to my solo consulting practice.  I’ve just read too many success stories (like this one) to not give it a try.


Levels of abstraction within AWS ML offerings

I went to an interesting talk last night at the AWS Boulder Denver Meetup where Brett Mitchell, the presenter, had built a sophisticated application using serverless and AWS machine learning technologies to make an animatronic parrot respond to images and speak phrases.  From what he said, the most difficult task was getting the hardware installed into the animatronic parrot–the speaker said the software side of things was prototyped in a night.

But he also said something that I wanted to capture and expand on.  AWS is all about providing building best of breed building blocks that you can pick and choose from to build your application.  Brett mentioned that there are four levels of AWS machine learning abstractions upon which you can build your application.  Here’s my take on what they are and when you should choose them.

  1. Low level compute, like EC2 (maybe even the bare metal offering, if performance is crucial).  At this level, AWS is providing elasticity in terms of compute, but you are entirely in control of the algorithms you implement, the language you use, and how to maintain availability.  With great control comes great responsibility.
  2. Library abstractions like MXNet, Spark ML, and Tensorflow.  You can either use a prepackaged AMI, Sagemaker or use Amazon EMR.  In either case you are leveraging existing libraries and building on top of them.  You still own most of the model management and service availability if you are using the AMI, but the other alternatives remove much of that management hassle.
  3. Managed general purpose supervised learning, which is Amazon Machine Learning (hey, I recorded a course about that!).  In this case, you don’t have access to the model or the algorithms–it’s a black box.  This is aimed at non machine learning folks who have a general problem: “I have structured data I need to make predictions against”.
  4. Focused product offerings like image recognition and speech recognition, among others.  These services may require you to provide some training data.  Here you do nothing except leverage the service.  You have to work within the limits of what the service provides, including performance and tuneability.  The tradeoff is you get scale and high availability for a great price (in both dollars and effort).

These options are all outlined nicely here.


Unintended consequences, vesting edition

I was having lunch with some folks yesterday and struck up a conversation with someone who’d been in the Bay Area for most of his career, but had recently moved back to Colorado.  We talked about hiring.  He mentioned that hiring in the Bay Area wasn’t hard, the real devil was retention.  As soon as one year had passed and 25% of an employee’s shares had vested, they were poached or off to the next startup.

Why?

Because from an employee’s point of view, 25% of the offered stock options for four companies is worth more to an individual than 100% of the offered stock options in one company.  It’s diversification 101. Most startups fail, and if you can quadruple your chances of owning options worth something, why wouldn’t you take it?  Even if you are leaving money on the table by losing the 75% of unvested shares.  (Of course, the cost of exercising the shares within the time limit matters too.)

This churn is horrible for the company.  The company spends two to three months bringing an employee up to speed and only get the remainder of the year of work from the employee.  When that employee departs, they take their experience and hard won domain knowledge.  A company can ameliorate this issue by making the job better (more responsibility, more opportunity, cool tech), adding to the diversity of the employee’s pay structure (paying more cash, which may reduce the need for option diversification) or increasing stock option compensation (more options).  The latter two increase labor costs, and the former is already emphasized in the initial recruitment pitch.

This is an interesting example of unintended consequences.  Two entirely reasonable premises: “we want our employees to be invested in the success of this fragile company–give them options instead of salary, but we want them to stick around before they get the entire payoff–make it vest over a period of years” lead to employees rationally making choices that hurt the company: “I’ll stick around for the year, but then I’m outta here”.


Publish java artifacts to s3 using maven

If you don’t want to run a maven repository server like Nessus, you can use AWS S3 for your maven repository for your java artifacts.  You can make the repository public in the same way as you’d make a website public.  However, it’s more likely that you’ll want to make it private and authenticate using AWS IAM.  Here are some step by step instructions that appear useful.

(Note there’s another S3 maven wagon, but it does not appear to support authentication.)

Why do this?  It’s one more piece of infrastructure that you don’t have to maintain, update, and run.

 


Book Review: Beforelife

I really enjoyed Beforelife. It’s a whimsical stroll through a fantasy land that exists after death.  The main character, Ian, arrives in the afterlife with an unusual trait: perfect recollection of his life before death.  This leads him to be institutionalized, where he meets some madcap associates, including multiple people who think they are Napoleon and someone who thinks he was a dog.

The rest of the book is all about how this group helps Ian discover what’s special about him.

I enjoyed the plot twists, which I won’t outline because spoilers.  There are several other historical characters, some named (Socrates, Isaac Newton), others hinted at.  I also enjoyed the world construction.  The world after death is populated by immortals, with interesting ramifications for areas of study like science and social constructs like policing.

If you’re looking for a light hearted jaunt and you liked Douglas Adams or Terry Pratchett, you’ll enjoy Beforelife.



© Moore Consulting, 2003-2017 +