| Episodes | 83 |
| Mentions | 145 |
| Cited here | 65 |
| First — last | #22 — #670 |
| Top guests | Elecia White, Henry Ott, Scott Driscoll |
| Related | firmware · sensor · bluetooth low energy · engineering career · product development |
Consulting in electronics is the sale of engineering expertise to client organisations on a temporary basis, either as an independent practitioner or through a firm. An hourly consulting business is fundamentally the sale of the practitioner’s own time, which caps revenue at hours available multiplied by rate.[144] It functions as a structural workaround to the management-versus-technical-track problem in engineering careers, because expertise can be rented at a premium for a short run instead of being bought as a permanent hire, and engineers have used it to return to technical work from management.[569] Independent consulting at current scale is only feasible because widely available networks and mature technology stacks let a small operator deliver what previously required a company.[645]
Terminology
One working distinction holds that contracting is doing the implementation work while consulting is planning with managers, doing architecture without implementation, mentoring and running reviews, though the terms are otherwise used interchangeably.[422] Contracting in the recruitment-agency sense means being placed by an agency and embedded inside a client company for a defined term of roughly six to twelve months to work on one project, which differs structurally from independent consulting.[541] Such placements are largely brokered through larger firms, a different commercial arrangement from being engaged directly as a named subject-matter expert.[421] Fast-growing companies also hire consultants to work on site, which can mean relocation and effectively functioning as staff while remaining independent.[424]
A workable test of professional-level competence in a discipline is whether someone would be engaged as a paid consultant specifically for that one piece of knowledge, rather than whether they have completed one or two successful projects.[447]
Pricing and billing structures
Engineering consultants price work either as an hourly rate or as a fixed bid, and the hourly rate is treated as negotiable, often set lower where the consultant has less prior experience in the specific subject area.[201] An hours-based hardware bid is typically framed as an estimated hour count at a stated rate, for example forty hours at one hundred dollars per hour, with the fixed job bid as the alternative structure.[384] Selling a solution to a defined problem earns more than selling hours, so pricing by outcome rather than by estimated effort is the higher-yield structure available.[384]
The billing model changes client expectations as much as it changes risk, since hourly billing carries an implied obligation of speed and efficiency while a job price shifts the client to a done-when-it-is-done posture.[201] Some consultants move well-understood work onto fixed bids to remove the self-monitoring burden of hourly billing, where any lapse in full concentration during a task such as PCB layout produces guilt about the clock.[201] A widely circulated consulting essay advises against quoting whole jobs where an hourly arrangement is possible, on the grounds that a whole-job quote places all schedule and technical risk on the consultant.[70]
That risk is concrete. Under a fixed bid the effective hourly rate falls monotonically with hours spent, so on a five thousand dollar job the realised rate is determined by how well everything goes.[404] A first fixed-bid job by a new consultant overran the quoted hours by a factor of 2.5, illustrating that estimation error on unfamiliar work falls entirely on the consultant under a fixed price.[404] On a fixed-price engagement an erratum in a chosen silicon device is absorbed entirely by the consultant, since the extra debugging effort is unbillable and the payment is all-or-nothing.[70] Mature, widely used toolchains carry lower latent-defect risk than new parts, because far more engineers exercise a mainstream C++ compiler than a newly released embedded device, so defects surface and are fixed earlier.[70] Time lost to defects in a consultant’s own development tools is not chargeable to the client, who treats the toolchain as the consultant’s problem, so tool risk is an unpriced cost of the engagement.[404] Clients also apply incremental scope pressure with small changes that a written contract does not fully prevent, because refusing each one damages the relationship.[404]
Setting a rate
Consulting rates track local cost of living, so quoted rates from high-cost regions such as California run substantially above those elsewhere, which surprises clients who have only benchmarked against salaries.[70] A rate has to cover self-paid insurance and taxes and additionally the unbilled time that surrounds the work, which is why the headline number sits well above an equivalent salary.[492] Rates look high to clients who compare them to salaries because they omit what the client is not paying for: insurance, benefits, and the obligation of continued employment, since a consultant can be released immediately.[409] Moving from salaried employment to consulting is described as roughly tripling the headline rate while gaining control over which hours are worked.[264] In the United Kingdom consultants earn considerably more than engineers in salaried positions, whereas the US comparison is complicated because healthcare and similar costs are bundled into employment.[549]
Undercutting pressure is a direct function of substitutability: a task a hundred other consultants can do equally well invites a price war, whereas specialisation and being wanted in particular insulate the rate.[70] Chronic underpricing is a specific form of undervaluing one’s work, and in practice rates are raised only when demand physically exceeds available hours rather than on any deliberate schedule.[349] The demand signal that forces a rate rise is concrete: requests amounting to about 80 hours of time against roughly 40 billable hours available, which itself represents around 60 hours of actual work.[349] Quoting a deliberately high deterrent price is a failure mode as well as a filter, because a client may accept and the consultant is then committed to work they did not want.[367]
Some handling of unfamiliar territory involves billing a discounted rate, or not billing at all, for the portion of time spent learning, and charging full rate only for time spent producing.[492] Part-shortage rework is separately uncomfortable to bill because the hours restore a design to where it already was rather than adding value, even though no mistake was made by the consultant.[601]
Utilisation
Billable hours are a fraction of hours worked: fifty hours spent on a project may yield only twenty to thirty billable hours, and that gap must be priced into the rate along with holidays and unbilled gaps between engagements.[70] Reaching forty billable hours in a week is rare in practice, and sustaining that level implies about sixty hours of actual work, since roughly fifty percent additional unbilled time is needed to keep on top of everything else.[409] Weeks with zero billable hours are a normal feature of consulting rather than a sign of failure, and experienced consultants stop reacting to them after enough cycles of forty to sixty hour billable weeks.[409] Self-employment does not automatically reduce working hours either, and consultants who set their own schedule frequently end up working 80-hour weeks.[22]
Early in a consultancy the self-teaching overhead is severe: on work sold as existing expertise, roughly ten hours of self-instruction can be required for every hour invoiced, a ratio that falls as accumulated experience grows.[560] Deep foundational learning, such as building custom logic with no datasheet behind it, is difficult to charge to a client, so part of consulting is unbilled time spent developing capability for one’s own account.[601] Doubt about being good enough is near-universal among practising consultants, and the work regularly requires learning something new either at the consultant’s own expense or, with agreement, on the client’s, with clients bringing genuinely unsolved problems considered the best clients.[549]
Because time carries an explicit hourly value, the buy-versus-build calculation changes across the whole business, and paying for software or services becomes rational wherever it frees hours that can be billed.[492] Consultants automate the back office accordingly, using accounting packages such as QuickBooks or Xero and paying a monthly payroll service that files taxes including awkward state filings.[492] Bookkeeping consumes enough of an independent engineer’s time that a bookkeeper is recommended among the first five hires of a growing consultancy.[422] Where consulting hours are expensive, paying premium fees for fast PCB and mechanical prototype turns is the cheaper option overall, since a five-day turn instead of a three-week turn removes consultant hours spent waiting and aligns with clients whose priority is time to market.[487] Concentration is protected in the same spirit, with social media clients and their notifications shut off entirely during billable work.[55]
Finding clients
Sourcing new clients, rather than doing the work, is repeatedly named as the hard part of consulting; most leads arrive by word of mouth and few consultants do any promotional work showing what they have built and how they operate.[404] Willingness to do sales is a precondition for going independent, since early income is lumpy, with jobs appearing and then disappearing, and a large share of early effort goes into finding work, establishing a name and identifying which clients value what is on offer.[305] The binding constraint when pitching a new consultancy is absence of a track record rather than the pitch itself, since prospects are being approached constantly and must extend trust without evidence.[635] Converting a former employer or past colleague into a first client is the ideal launch but is not reliably available, and more than one consultant has attempted it on leaving a job and failed.[305]
Client acquisition splits into push and pull: scheduled outbound contact pushes to a list, whereas publishing technical videos and articles pulls prospects toward the consultant, and both have produced paid engagements.[569] Systematic relationship maintenance generates work directly; one engineer used customer relationship management software to schedule contact, deliberately reaching out on people’s half-birthdays rather than birthdays so the contact stood out, and attributed jobs to that practice.[569] One documented route to a first client is direct outreach on a professional network after defining a target segment, in that case small and medium embedded and IoT companies, chosen because large companies were judged unapproachable.[635]
Bidding on open contractor marketplaces yields worse work than inbound enquiries generated by a known name, and the most effective way to build that name is to publish your own projects publicly until clients approach you.[324] Open online bidding marketplaces are dominated on price by offshore consulting companies, which routinely underbid individual local consultants.[83] General freelancer marketplaces such as Upwork, Fiverr and freelancer.com also give inconsistent results when subcontracting engineering work, since a job posting attracts a wide spread of applicants including unqualified ones.[527] Publishing reference designs against a newly available open platform is a route for a less established consultant to become the recognised implementer for that technology wave, worth more than the component discounts such work sometimes attracts.[517]
Formal channels exist alongside these. The IEEE operates consultant networks that provide a support and networking structure for independent engineering consultants.[41] Chip and software vendors publish lists of contractors competent in their tools or programs, with entry typically by application, and such listings are a lead channel that clients actively search.[119] Obsolescence work arrives through contract manufacturers and distributor field application engineers whose clients need an unbuyable part designed out.[409] Leaving a name in schematic comments or on production test fixtures makes an engineer traceable years after departure, and former employers use that plus professional networking sites to recruit them back as paid consultants.[108]
Specialising in a narrow technical area produces a self-reinforcing referral loop, since once a group of practitioners in that field know a consultant, they and their competitors keep calling.[160] Long-term business is sustained by repeat work and referrals rather than by maximising each invoice, so experienced consultants absorb small items such as a ten-minute clarifying phone call rather than billing them.[135] Demand also arrives in clusters, with the same class of problem appearing across several unrelated clients in the same year, which lets a consultant anticipate where expertise will be worth developing.[165] Writing up a recurring client problem where public information is scarce compounds into further work: a short web write-up on mixed-signal grounding led to a commissioned magazine article and then to a paid course.[165] Clients often prefer a local consultant in the same time zone not for sentiment but because escalation is faster and a failing project can be addressed face to face.[160] Naming a one-person consultancy something other than the founder’s own name is a deliberate positioning choice, though buyers scanning vendor listings can usually still tell an individual from a multi-person firm.[119]
Some engagements are closed off structurally. Certain companies deliberately refuse to contract out particular work, such as core algorithm development, because they want that capability retained in house.[422]
Conducting an engagement
Customer-supplied requirement documents are rarely complete or usable, so the standard opening step is for the consultant to write a specification back to the customer stating what they believe the requirements are.[135] Proposals should carry an explicit known-risks section, and stating up front that a particular task has never been done before and may fail is normal practice, with a good client accepting that framing and proceeding.[409] Calling an experienced practitioner for advice before accepting an unfamiliar job is a low-cost way to de-risk it, distinct from asking permission to take the work.[404] At a small, selectively staffed consultancy the individual engineers, not a manager, are expected to hold the schedule and cost conversations with the client and to be present when those decisions are made.[487] Every new client has to be convinced afresh that the consultant is worth the fee, and one deliberate technique is recounting selected past war stories in informal settings to demonstrate relevant expertise.[487]
Running several small jobs concurrently is only workable if customers have been trained not to expect unrealistic timescales; with realistic delivery dates agreed up front, engagements can be multiplexed.[135] Independent consultants do not escape having a boss, since each active client functions as one, so three concurrent engagements mean three sets of expectations to satisfy.[70] Being a client’s sole electronics resource creates an implicit duty of availability that constrains the ability to decline or defer that client’s work.[135]
Declining an unsuitable enquiry outright is cheaper than engaging with it, because once a prospect is given any opening they will negotiate and consume unbillable time.[285] Some engagements should be declined at any price for reputational reasons; an offer of roughly twenty to twenty-five thousand dollars plus travel to validate a purported energy device was refused rather than bid on.[530] Where work cannot be accepted, pivoting into an advisory role on that project preserves the relationship and the person’s access to the expertise without committing to the implementation.[670] Once inbound demand exceeds capacity a consultant can select only paid engagements, and unpaid work should never be accepted.[324]
Failure modes recur on the client side. A common one in advisory engagements is being given incomplete information about the situation, so that considerable time is spent producing recommendations that do not make sense against the real constraints.[422] Clients commonly assume a consultant will handle everything around the technical task, and consultants who possess only technical skills and execute exactly what they are told leave such clients exposed.[187] Client-side cancellation risk is abrupt and largely unmitigable: an engagement running for about a year ended when the owner simply decided to stop paying.[509]
Absences require explicit planning. Taking a holiday requires a policy decided in advance covering whether to go completely dark, what is communicated to clients, whether calls are answered, and what counts as an emergency.[413] Planned extended absences are handled by giving clients around six months of notice and sequencing projects to complete or pause cleanly, with the real reason stated rather than a vague unavailability.[541]
Starting a practice
The recommended first step is not technical but financial: clearing debt and putting personal finances in order before starting, because that is what carries the practitioner through lean periods.[409] The recommended cash runway is around twelve months of expenses, adjusted for household circumstances since a partner’s full-time income materially reduces the buffer a solo consultant would need.[409] In practice a consultant setting up alone often starts with far less, in one account about two months with no interested customers, which is why the first period is dominated by cash anxiety rather than technical work.[560] A lower-risk route out of employment is to leave only once an independent recurring revenue stream plus a visible consulting pipeline already exist, with the runway limit understood in advance.[184] Starting young lowers the required income floor; one practitioner took several years across two ventures before earning reasonable money and credits the ability to run lean during that period.[409]
The business skills that must be acquired in parallel with the technical work are quoting, client management, deciding what is and is not billable, and accounting and invoicing.[404] The non-technical skills also include how to pitch for work and how to report completed work back to a client, typically learned from practitioners already doing it rather than from technical training.[349] Years spent as an applications engineer managing customer relationships, including delivering the news that a requirement cannot be met while preserving the long-term relationship, is described as good grounding for later consultancy work.[452] Prior experience managing a team, while not a prerequisite, materially helps a consultant who begins hiring subcontractors.[527]
A practical signal that an engineer should go independent is doing the technical work they want unpaid in their own time, for example staying up until three in the morning doing PCB design while the paid role has drifted toward marketing.[549] Absence of a formal degree closes some salaried routes while leaving consulting open, so several self-taught engineers have used independent work as the route into technical careers that employers would not offer them.[374] A common bootstrapping pattern is to take on roughly twenty hours per week of consulting purely to cover living costs while the remaining time goes to an independent product or research effort.[183] In a downturn when companies move to cash and liquidate, part-time consulting in the sectors that expand is a hedge available to engineers whose main work has contracted.[486]
Capital equipment failure is an immediate, unbudgeted cost for an independent: with no employer to supply a replacement, a failed laptop had to be replaced the same day at around one thousand dollars to keep working.[409]
Working conditions
Independent consulting removes the organisational guardrails that a company supplies through established process, so responsibility for every check, including elementary ones, rests entirely with the consultant and nobody will ask for the missing step.[584] With no colleague to check the work, a consultant must become their own safety net, which extends to learning each new certification or specification from the written standard and its nuances when a project requires it.[588] Self-review has a ceiling that diligence cannot overcome, captured in the mantra that two people working as a team are worth three, because differing perspectives challenge interpretations that a lone reviewer will not question.[588] A third-party design review paid for before fabrication generally costs less time than the debugging it prevents, and teams large enough to run internal reviews should make the review plus an explicit test plan a required step.[584] Where a design engineer inside a company can escalate through distributor representatives, field applications engineers and factory applications teams, an independent consultant reaches the end of that chain and must resolve the problem themselves.[452]
Isolation is a recognised working condition of solo consulting, and some practitioners run background audio through the working day to substitute for ambient office conversation.[176] The largest reported shock on leaving employment is not the loss of the paycheque but the loss of colleague camaraderie and the ability to talk a problem through informally over lunch.[409] A semi-private, application-gated peer forum for consultants exists so that practitioners will discuss client and pricing details they would not post publicly, while making repeatedly asked questions searchable.[479] Community advice differs in kind from individual advice, returning contradictory experiences, and its value lies in the cases where independent accounts reinforce each other, which is the signal that a pattern is real.[479]
Consulting is attractive partly because it concentrates time on design and reduces exposure to documentation, regulatory work and repetitive testing that many engineers do not enjoy, alongside schedule flexibility.[549] It exposes an engineer to a wider range of problems than a single full-time role, which counters the skill stagnation that follows from repeating one class of design, at the cost of having to learn unfamiliar domains quickly on a client’s schedule.[422] It suits engineers who change context frequently, because serving successive clients means moving between different problems and technologies by definition rather than as a career flaw.[635] A stated attraction of design-for-hire is the externally imposed constraint set, since self-directed projects tend to be given constraints that are never realistic enough to force real engineering decisions.[675] Against that, enjoying personal engineering projects does not predict enjoying engineering for a client, and some practitioners find the obligation to build someone else’s design a persistent demotivator despite liking the technical work itself.[675]
Consultants also become pigeonholed because risk aversion is rational for them: given a choice they take work resembling jobs already completed over unfamiliar work that looks more interesting.[406] Investments such as making a design modular against part shortages do not pay back for a consultant, because each engagement is a substantially different design every few months and the reuse never accumulates.[541] Where a design carries no product differentiation in the component choice, some consultants select parts solely by existing driver support in their chosen framework, declining any device without a Zephyr driver rather than writing one.[645]
Scaling and diversification
A pure consulting business has only two scaling levers, working more hours or charging more, and both are bounded, which is the standard argument for adding developed products alongside client work.[232] Because a consultancy sells a fixed stock of time, its only scaling routes are hiring people or becoming markedly more efficient, whereas a product business scales differently by adding automation and staff to something reproducible.[511] A solo consultant’s throughput ceiling is headcount, and work is routinely turned away not for lack of interest but for lack of people to execute it.[303] A consulting firm working roughly ten concurrent projects lacks the slack to train a graduate engineer, so hiring favours established experts who can be productive immediately.[624]
Established electronics consultants commonly run at least one revenue line besides client work, such as products, a seasonal engineering programme or a paid course, which raises the probability of covering income in any given period.[409] After about five years a consultancy accumulates two distinct buffers: cash to bridge periods without customers, and a reusable library of intellectual property such as DSP algorithms, frameworks and known platforms that makes new work quotable with less risk.[560] Process discipline that a client insists on can be repackaged as a marketing differentiator when pitching subsequent work.[627]
The relationship between products and services takes two structures. One is a single company that both develops the owner’s own products and sells consulting services, offering either complete device development packages or project-based work.[409] The other keeps the two lines separate, with a consultancy of about ten years’ standing running a product line with almost no crossover between the product customers and the consulting clients.[409] Running both can be mutually reinforcing when the product line serves the same niche as the client work, which is a reason to retain the services arm rather than exit it.[226] A repeated client request is the signal to productise: when many clients ask for the same building block, a standard module can be sold to those who can afford neither the consultant’s hours nor a fully custom design.[226] Hardware sold into a technical workflow is often bundled with consulting on its use, so the offering is a package of device plus the expertise to apply it.[369] Companies attached to maker and developer communities monetise that reach the same way, selling consulting services around the community’s hardware.[350]
The tension runs the other way as well. Consulting work has kept small product companies solvent for years, but every engineer assigned to client work is taken off the company’s own product.[458] Consulting revenue arriving as intermittent capital injections can also slow a product business, because each new client payment removes the urgency that forces the product forward.[280] Funding a hardware product company out of consulting revenue is nonetheless a well-trodden path, followed among others by the founders of Punch Through and of Saleae in their early years.[280] For a small independent hardware engineer, consulting can return more money than selling a self-designed product, because product margins at low volume are extremely thin.[153] At one-off and low volumes engineering effort dominates the bill of materials: a device with roughly one hundred dollars of parts can require about five thousand dollars of engineering time to reach working condition, though the second unit then costs only the materials.[422]
Adjacent service businesses face their own structural limits. The US contract manufacturing landscape splits between shops built around ISO and aerospace or military quality standards and smaller maker-oriented shops without such certification; competing as a general contract manufacturer means fighting on pennies, and volume production tends to move offshore regardless.[458]
Relative to employment
Compared with employment, independent consulting raises both the upside and the downside of the same career, and the practitioner either accepts that variance or does not start.[492] Over a ten-year horizon consulting income does not automatically beat the salaried alternative; one consultant calculated that accepting a contract-to-staff offer a decade earlier, even without raises, would have left him about the same or better off.[409] The security advantage of employment is partly illusory, since whole engineering divisions were eliminated during the 2008 financial crisis and employees could arrive to find the job gone.[409] Deep single-topic specialists inside companies are structurally exposed in the same way, because the role disappears when that specialty is no longer required or can be sourced more cheaply, which is one argument for holding expertise as an independent.[307]
Formally agreed reduced-hours arrangements inside companies are frequently penalised at performance review, with employees marked down for not matching full-time colleagues despite the agreement.[264] Engineers also frequently absorb responsibility for failing projects that belongs to management; at one firm the entire engineering staff resigned after one or two years of 80-hour weeks and heavy travel when it became clear that effort could not compensate for repeated bad decisions above them.[183] Running a consultancy in turn teaches sales, business development and the cost of sales, so an engineer who later returns to a corporate applications or design role brings back an understanding of the commercial side they did not previously have.[492]
Demand for consultants tracks conditions in the wider industry. During the semiconductor shortage consultants were uniformly busy because clients needed help redesigning around unavailable parts, and the shortage coincided with a labour market in which hiring a full-time engineer was effectively impossible.[601]
Authorship as an adjacent activity
Measured per hour, writing a technical book returns far less than consulting at the same author’s rates; its value lies in doors opened rather than royalties, and one author traced an entire subsequent hardware company to the book.[123] For technical authors the book itself is a loss-making project at contracting rates, and its economic function is credibility that generates teaching engagements, with one embedded-systems author deriving about 90 percent of income from teaching rather than consulting.[187]
Project execution practices
General advice for installation-scale work is to plan for every contingency, layering backup plans at the design level rather than only on the day of installation, for instance ensuring that half the light strings still function if the rest fail.[404]