Electronics Product Management
It’s No Country for Old Men.
If you spend significant time reading Medium, Hacker News or similar blogs, you’d be forgiven for thinking that all you need to succeed as a product manager is a Lean Startup mindset, a few growth hacking tricks, an Optimizely account… and you’re off to the races.
I’d like to describe a different world — the world of electronics product management. It’s not for Old Men.
Enterprise electronics customers are really, really pragmatic people. The name of the game is risk reduction.
I have a Dropbox folder stuffed full with articles on behavioral biases, bracketed pricing tricks, viral marketing tactics, advanced analytics, and so on. They are fun to read. Honestly, most are not all that useful.
Your shiny new electronics thingy is going to be evaluated by a minimum of 3-6 people, each with their own criteria. Engineering organizations are fundamentally conservative and HATE risk. Your best bet is to deliver a clear, easy-to-understand value proposition with a minimum of fluff. (In other words, don’t worry about whether you’re using a “.99" in your pricing. It won’t matter.)
Product definitions: Same degree of difficulty. Just different.
Enterprise customers often — but not always — have firm ideas about new product features. If you’re lucky, they are written down and provided upon request. 50% will be “must have” or “nice to have”, but expressed verbally.
Other customers won’t provide any firm guidelines, but instead will go with “the industry standard”. Fortunately the electronics world does a good job of defining standards.
The biggest risk is of being completely wrong on a hardware feature that goes into production. The cost of a retrofit is 1) exponentially higher, and 2) not just a matter of pushing a new code commit into GitHub. It takes weeks, if not months, to fix those ulcers.
Running multiple sales funnels and/or cohorts is a LOT harder.
First: you’re probably in charge of not just a single app, but multiple products. (I have ~25 on my plate.) Each has distinct timelines, financials, value propositions, and feature sets. Funnel management is doable — but you’re duplicating the effort 25 times.
Second: Sales leads come in from multiple directions. Chip companies use a combination of inbound marketing, direct salespeople, manufacturer’s reps and distributors to service their customers. Each channel has its own sales funnel, with unique CRM tools | datatypes | workflows. A mastery of Excel is usually enough to wade through the inconsistencies. But data munging is not a good use of your time.
Third (and this is the most intractable): you’re dealing with the hidden motivations of salespeople who don’t report to you. Each sales lead score is subjective & situation-unique. And you won’t get Win/Loss data with believable root causes unless you take the buyer out for drinks. Without the salesperson around.
The financials are way different.
One thing that I’ve noticed in SaaS-oriented blogs is that nobody seems to talk about “product cost” once something is released into production. Cost of customer acquisition (CAC) is one thing — and it’s an obviously valid metric. But it’s not the only thing. Not even close.
In the semiconductor world, it’s easy to spend $1-5M on upfront development costs before you have a final device. A maskset, engineering wafers, packaging materials and tester time can add another $500K to the bill before you receive your first product samples. And every part that ships afterwards will have an oh-so-real product cost thanks to materials and manufacturing yields. 60% manufacturing margins are great if you can get them. 40% is much more common. And this is before your friendly CFO loads you up with corporate cost allocations.
You can — and should — look at your customer base with CAC and LTV metrics in mind. But those product costs will kill you if you’re not careful.
Your shiny new thingy is merely a component in a larger system. Integration, integration, integration.
Let’s say that you’ve successfully run the gauntlet of uncompromising designers, paranoid manufacturing quality engineers, and the just-give-me-the-price-and-part number buyers. Everyone’s satisfied, so your customer decides to order 1000 parts for a beta manufacturing run.
This is no-man’s land, because what looks great in simulation can often go awry on a real PC board. I’ve seen opportunities go belly-up due to mechanical defects, thermal issues, and signal integrity genies. You just can’t predict them. And the customer is just as upset because they’ve wasted $50K on a production run just to qualify your device. Not good.
Product ramps are lumpier than my grandmother’s mashed potatoes.
The customer calls to tell you that the beta manufacturing run was successful. You’re “qualified” and “designed in”. Pop a cork!
No, young Padawan. Not even close.
That forecast that you get from the buyer? At 6 months, that forecast will be off by +/-50%. 12 months? Bust out a Ouija board. Plus - the timing of the ramp is entirely dependent on the success of the customer’s end product. The volume may be accurate — but may be 9 months late due to some other component’s availability. You simply can’t predict it.
If you’re lucky, your device is the only possible option for the customer’s project at this point. (Think microprocessor.) If not (memory chips, for example) then you’re probably competing with one or more other suppliers for market share. 1,000 parts/month? That’s just your share — which is a function of price, service, and quality. Not the total.
The customer: “You’re the man. You’re my partner. You’re STRATEGIC to me.” The customer’s 3rd-party manufacturing partner, on the other side of the planet: “Please provide your best price and lead time for the following PN#, for anonymous new customer project.”
Different organizations with different agendas. If you’re lucky the customer will resolve these situations for you.
A previously unknown customer: “Hey! I purchased some of your parts from (an unauthorized broker), and they don’t work! WTH?”
It’s pretty common for counterfeit chips to appear in the electronics supply chain. They’re like bedbugs — you just have to deal with them when they appear.
Don’t get me wrong — electronics product management can be very rewarding.
But the variety of screwups along the way will change your worldview if you’re not careful.
I hope you enjoyed this gentle rant. If so, let me know at @brianpiercy. Thanks!
