Tuesday, September 24
I’m at my desk when a little bell sound emits from my phone. It’s a slack notification from the Senior Hardware Engineer for a startup in Cambridge. We’ll call him “B”. In Fall of 2019 I was working as an independent contractor, B had brought me on to help get their devices out the door. A good customer. The first of these devices we would reference as “the buttons.”
[B] – How are the devices looking? Do you have an estimate as to when they’ll be finished?
[Alex] – I’ve been pulled away from the buttons for the last week. I’ve had a job getting a prototype ready for a MassChallenge participant. Their pitch is tomorrow, so today is crunch day – but I think I’ll get it working.
I plan on returning to the buttons post-facto, most likely today. I’d like to get a little more time with all the steps in order to give an estimate. Can I get back to you later today, or worse – tomorrow at lunch with an informed estimate?
[B] – Our target was to ship this week, but let me know if you can finish them this week to ship early next week.
That was news to me, or at least a shock. The ship date had been pushed back because of delayed sub-component shipping. We’d already passed the original ship date. I wasn’t sure where this new one had emerged from. Perhaps it wasn’t communicated, perhaps I missed it due to the horse-blinders of the Mass Challenge job. It didn’t matter much, as it was. I had one of two choices:  go back into the contract and try to be particular about what had been scheduled or  do my best to meet this customer’s timeframe. This was a good customer, and I wanted to continue doing business with them, so choice  it was.
[Alex] – Ok, I’ll get back to you with an informed estimate.
Now, I had been tracking my time up to this point. My work as an independent contract this year had demanded as much for invoicing. I was using Google Sheets and basing it off my experience with tracking my sleep for the past 3 years. It was simple, but effective. I would list a start time, an end time (from which a total could be calculated), and notes for what was done during that period. I would do my best to break up and record sessions for individual tasks. Everything was down to a 15 minute resolution.
I could see with blatant transparency that I had not worked on this commission in the past 6 days. I knew why, it was the aforementioned Mass Challenge commission that suddenly took on more urgent work. I had consciously put this commission on the back burner, knowing that I would have enough time to address it once the Mass Challenge one was done.
Or so I thought.
With B’s message to me that they had planned on shipping this week, it was a slap in the face. I knew I wanted to do as well as I could for them.
Having made the decision to finish this device commission as soon as I could, I again found myself with two options: [X] I could say “yes, I can have them finished this week” like he was asking and just put my nose to the grindstone to make it happen or [Y] I could take another approach – an approach I had been looking forward to taking. Indeed, one of the main reasons I had taken on this commission in the first place. I wanted to get an informed time projection.
A little bit more about this job vs. the work I typically do. Since having left a giant aerospace company earlier this year I have been taking on commissions out of a makerspace called Artisan’s Asylum. Most of them have been for startups, a customer and culture I find rewarding to work with. Most of them have either been debugging electromechanical systems or designing new ones. Those jobs have a lot of uncertainty. In debugging, the uncertainty is even hard to describe. Let’s just say I have a new respect for auto mechanics being able to quote the hours needed for a repair – although it is often wrong. In design work, you are creating something new. It’s hard to create bounds for that.
This job, however, was different. This job was–well–boring. A local startup was creating about 500 units of their product in their first small batch. It was about 300 units of reconfigured Amazon IoT (Internet of Things) buttons and about 200 units of a custom PCB (printed circuit board) and custom enclosure (to house the circuit board). The kicker, everything was already designed. Indeed everything was ordered and sourced and had a shipping schedule. An engineer had already done the “heavy lifting” – the part I usually do. All that was left was assembly-line-esk work – something I am not used to doing. I was going to be repeating manual tasks of putting pieces together or configuring electronic devices on the web.
Mind numbing work.
However, I had a taste of this when I had created 64 units with an electrical engineer in the summer earlier this year. We had designed and assembled that together. That’s where I got my first taste of setting reasonable specifications for a product and understanding the accuracy deviations manufacturing processes. Learnings very important for the design side of production. However, there was the assembly side of it as well.
Scaling my efforts is important to me as a person – I know this about myself. I want my designs to help many, many people. That means understanding the production process. That means understanding what it means to assemble a design.
Now, this is at the crux of my thinking and reasoning for taking on this job with B. Sure, I could take the same path some of my entrepreneur friends have taken – come up with a “looks like” idea on the back of a napkin, fly to China, and talk with engineers there in a production house to see how much money and time it would take. But that, to me, is a black box. One I don’t have insight or control over. I want to build up my understanding from first principles. And in this case, that meant doing it myself.
I’m very glad I did.
I decided to take a different path than just “yes, sir! I’ll get them done this week like you ask. It’s only Tuesday, and there are only 300 devices. I can just pull enough late nights to get that done, right?” That is, essentially, the same approach that many college students use when it comes to writing papers or studying for examples. Other students. Me? No no no. I was… certainly, not one… no, not me… no…
Yes, that was me. That’s a picture of 20-year-old Alex studying for a test the night before. The timestamp says 1:19AM, and I remember being more than a few hours into studying. I took that picture to remind my future self how much it sucks to not be in control of time. I know that I didn’t have a good relationship with time in college. And after, honestly. As a person who wants to scale their efforts, time has to be understood. I wanted to change.
Plus, I couldn’t make that uninformed promise this time to B. I had spent too many 3AM nights working on this Mass Challenge commission. I needed to rest. I also had a friend staying a day in Boston this week that would take me away from work for an evening and a day.
I needed to plan. Moreso, I needed to project.
This is what I had been seeking. Now, circumstance and need were forcing it to happen. Again, I’m very, very glad it did.
I knew that these buttons had a set number of steps in order to finish them. Eight steps, as it would turn out. Up to this point, however, I had only completed two of them. I needed to get through all eight before I could begin to even fathom creating a projection for the whole project.
This is a vital step.
If you have not been through all steps, you have not created the boundaries for the project. This was a lesson I–fortunately–learned early in my entrepreneurial experience. I have seen myself bettered by following this doctrine, and I have seen others fall because they have not.
This is so vital that I’m going to repeat it. Finish an entire unit before you attempt to scale it. I could write a blog post five-times this one’s length on that matter. Indeed, many have. I would direct you towards the book “Lean Startup” by Eric Ries to learn more.
For the buttons, my plan was set. I would finish a small number of units completely so that I fully understood all steps, any surprises would be flushed out into the open, and I could organize the steps into a system. Then, and only then, I would have a list of those needed steps and the time projected to complete them.
Wednesday, September 25
My plan was set – time to execute. As I went through the unknown steps I collected the data I would need for my projections. I was more detailed in my logging than I had been previously. Now I was taking measurements down to one minute resolution. I chose 15 units as my sample size and timed how long it took to complete each step. 30 would be better, but I was short on time and I made that call. I would normalize that total time per step to get a single unit time. Lastly, knowing how many units were still subject to each step (a different number per step, I will add), I would linearly interpolate into the future. I would get an estimate of how much time would be needed for all steps to see all buttons ready for shipping.
I was going to get an informed prediction. I was so excited!
Let’s sum it all up, and there we have it! 27 hours! Well then!
I felt so proud. I felt so informed. I felt in control.
Now, I knew from past experiences that little hiccups can occur. I needed to give myself a little breathing room, so I added 3 hours to the total. 30 hours.
Neat right? Simple in theory, but it has been a challenge all these years to execute. It’s literally just using algebra we learn in middle school. At least this version. Trust me, it can get more complicated. That is something called a “controller” which is used largely in robotics and complicated engineering systems. That can use calculus to understand the past and predict the future. That was a little outside of our range here though.
It was Wednesday afternoon, I needed to get a response to B.
[Alex] – Ok, I have an informed estimate for you. It’s based on the most efficient route I’ve found (big time savers in packaging and router configuration). To finish up to unit number 350 will take 30 more hours.
[B] – Gotcha, when do you think you can have the order completed?
[Alex] – I plan the earliest I can finish them is end of Monday. That’s having everything ID’ed in a spreadsheet, labeled, configured, claimed, and packaged.
Does having them finished Monday end of day work for you?
[B] – Yes, that works. If you’re confident in that date, I’ll let our operations team know to prepare documents for freight shipment.
I breathed a sigh of relief. I was glad to hear the push back was acceptable. Through my projections, I learned that I could not finish by the end of this week. Option [X] above to (blindly) promise what the customer wanted and figure it out after the fact was not reasonable. Had I taken that route, it would have lead to even more stressful late nights. Here’s a little secret through: having made these projections and communicated them to the customer, there was no stress. All expectations were set and reasonable.
I had my hours planned out for the remaining days. I would work 10 hours Thursday, enjoy that evening with my friend, enjoy Friday with my friend, work 5 hours both Saturday and Sunday, then finish with 10 hours on Monday. I stuck to the schedule and communicated with the customer as I went. I, of course, continued to track my time. Not to the resolution I was earlier, however. That would have been interesting from a “learning curve” perspective, to see if certain tasks got faster. Something for another time.
In the end, nothing was left unaccounted for. I even took the time to write a detailed 50 step list of all the steps in sequential order. This assembly line could easily be picked up by myself or someone else, and all my efficiencies would be captured.
So… how did it all end? Is this the part of the story where I say some bug in the system, some dragon reared its ugly head and sent me into a tailspin? That only by my grit and determination did I get things done on time? Did a friend come through in my time of need and help me finish? Did the excel version of Excalibur get hurled at me from some lady in a lake?
No. Everything went to plan.
And it was a joy.
I spent an evening and a day with my visiting friend without a worry about the project looming in the background. I had no more 3AM nights. I had no worries about straining the relationship with the customer. All of that was mitigated by my projection into the future. Turns out my 15 samples was enough to make an accurate projection.
Guess how long it took. I originally projected 27 hours and added 3 for buffer. 30 hours total. Go ahead, take a moment to guess.
After the time I spent on that Wednesday to make the projections, it took 29 hours to finish. I was right on the money. I was within an hour. That feels great. That gave me confidence going forward.
It would turn out that I would need that confidence a few weeks later. Remember how I said there was a second device of 200 that needed to ship as well? That assembly line was arguably much more complicated. However, I was tracking my time from the beginning given this new process I had practiced.
Now, I wasn’t the lead. I wasn’t responsible for the ship dates or allocation of resources. B was. But I very much had plans to be responsible for those aspects in the future for my own projects. This was my chance to learn.
I had worked a number of days in B’s office assembling the parts. At the end of the week things were getting a bit hectic. B had set the ship day as end of business day Monday. It was now Friday afternoon. We had some done, yes – but were we on track?
I took a couple hours Saturday morning of my own volition to calculate. I had been–again–entering in the start and end times as I went. Granted, this time there was no sample size of 15. The quantities were all over the place – I was just working.
I divided out the notes into tasks, just like last time, and normalized the time to complete each step.
Unfortunately, I didn’t have the opportunity to completely finish all units on this one. You’ll notice that the last step “attach back panel” has a purple highlight on it. I was not able to finish this task because attaching the back panel would make the boards inaccessible and B still needed to access them. Now, if I had been planning to do a full time projection I would have planned to do this step at least 15 times. Even if it had to be undone, I would then have a fact based set of data for projecting that step. As it was, this Saturday morning, the best I could do was relate it to a very similar steps of “screw in the boards”. I copied that projection as it stood. I didn’t want to leave a “known step” unaccounted for.
So, 15 hours. I don’t know about you, but I didn’t want to work 15 hours on Monday to get everything done and out that day. That’s even planning ahead, let alone being caught off guard at 5PM, seeing that there is much work left to do. I penned an email to B.
I hope you’re enjoying the weekend. Not to bring up work (but to bring up work), one thing I have been practicing lately is interpolating the time to complete tasks on a small subset of units to the rest of the batch. I calculated that this morning, and project that there is at least another 15 hours of work to get all 225 units complete.
My projection does not include individual boxing, testing rig time, or loading into the large box. Maybe worth me packaging up 10 of them individually first thing Monday to get a projection. All said and done, probably ~20 hours of work left.
I’m available Tuesday as well if needed, but not Wednesday.
Here’s a screenshot of the final calculations, taken from my timesheet.
I chose to be a little tactful in my message. Instead of calling out his ship date as unattainable, I just relayed the required man hours to complete it. Who knows, he may have been able to call in help.
When I arrived to work Monday, I learned that he had requested a push back with Ops to ship on Wednesday. That felt pretty good. As for my projection? I had a hard calculation of 15 hours and added an additional 5 hours for a total of 20. Go ahead, take another guess. How do you think I did?
I worked an additional 19 hours and 15 minutes. Again, I was within an hour. And B shipped Wednesday. Mission accomplished.
I felt great. I felt in control. I wanted to practice this more and help spread it to others. I’ve started offering “time calibration” consulting. I’ll sit down with a project lead and go through a project that… well, maybe didn’t go according to plan. We’ll do our best to retroactively collect all time data that we can to give them a picture of what happened. This will often involve money spent on the project as well. The results?
I happened to capture two pictures of epiphanies in-the-moment. It was actually pretty cool. The first time I helped someone with their calibration was another independent contractor who had a project that went way too long and had way too many late nights. As it was, it was still three weeks behind the original ship date. Through our calibration, we learned that the original projection for the project was 80 hours, but it ended up taking 500.
Most notably, we learned that there was a single feature that accounted for 150 of those hours. It was a feature that creeped in during the end of the first meeting with the customer. Later, as she was communicating with the customer about schedule push back, cutting this feature was never brought up. She didn’t know it was the culprit. Had that been known as an option, and the customer agreed (which they most certainly would have), a week could have been shaved off already extended three week push back. Just project wise, that would have made all the difference (let alone an emotional difference for those toiling away on it).
This is the contractor’s reaction when she came to that realization.
With her eyes up-and-to-the-left in thought, with a subtle smile on her face, she airlessly mouthed a four letter word I’m sure we’re all familiar with.
She continued to sit there in thought. Contemplating what could have been different. Bare in mind this is the day after handing off the project. The wounds were still fresh.
“Now I’m just pissed… I should have done that.”
And like that, I converted her. She’s looking at time calibration in a new light now. She’s looking at this project in a new light too. It very much still sucks, but there was a major learning now that could be brought out of the pain.
I worked with another individual to calibrate their perspective of the time and money spent on a kickstarter they launched. These revelations are on the other end of the spectrum – quite happy ones.
Over 10 hours, we combed through pictures, receipts, and instagram posts relevant to his kickstarter campaign. I asked questions along the way, having him to guess how many hours he spent and how much money. We quickly learned that everything he remembered were the parts that really sucked. He remembered the project in a negative light because of this, it wasn’t fair to himself. There was much more work done well than he remembered doing. He also gained great insights into his motivations for working and how he operates. That was an unexpected pattern we recognized.
Most importantly though, we took a look at the finances of the project. He thought it was a wash. He thought he broke even or actually lost money. Turns out, he didn’t. He actually did quite well. We broke down his expenses into categories using my learning from my own financial tracker (have I not written about that? It’ll blow Mint and QuickBooks out of the water). We got a “cost of goods” for each of the production units made. He was within 97% of his initial project of unit cost from his original bill of materials. Some costs went up, some others went down, but ultimately he was right on the money.
He was ecstatic, and I don’t use that word lightly. Here was a year-long project and his first kickstarter–something he thought was “a wash”–that was actually a success. By digging into the time and money of the project in retrospective, he saw by the numbers that showed that to be true. It was only the pain of the project and he not having recorded his numbers that got in the way of seeing that.
“Dude, I made money! This makes me feel so much better. *laughing* This makes this kickster not a bust.”
Nothing about the kickstarter project changed, only his knowledge and perspective of what happened. It was quite incredible to see how that affected him.
Now he’s making plans to take the kickstarter product into small batch production. Injection molds and everything. He gets to level up into small batch manufacturing, which has been a goal of his. He’s also taking quite a different approach to time management going forward with projects. I don’t think it’s hyperbole to say that our time calibration has altered the path of his life.
I know that I want to create products that improve people’s lives, at scale. That isn’t possible without real time management and is something now I have created a habit of doing. Maybe this is something taught only to production engineers in grad school or on the job – but it is something we all can do.
If you have had a project that went off the rails or aren’t sure how to measure the success of an initial product offering, let’s talk. I’ve seen first hand how it can give you the control and confidence you need to tackle bigger projects going forward.