My Unvarnished Guide to Solution Engineering

Author: Max Halford Published: June 4, 2026 Tags: data-science, showerthoughts

Context

Halford joined Carbonfact (a climate startup) as its first employee, building software that ingests ERP and PLM data to produce environmental reports. He was pulled into sales and onboarding processes while also building the product. He describes himself as "awful at first," recalling harsh feedback from a cofounder that ultimately helped him grow.

Core Philosophy

He references Jim Donovan's 11 pieces of advice on client service as an "excellent starting point," then offers his own hard-won perspective as a solution engineer at a small startup rather than a Goldman Sachs executive.

The article organizes the customer lifecycle into three phases, each requiring a different attitude:

Phase 1: Presales

The prospect hasn't signed yet. The goal is to secure their commitment. Halford calls this the least stressful phase — "it's like flirting" — because no one owes each other anything yet. He emphasizes that prospects are anxious about choosing correctly, since picking wrong "could cost them their job."

Understand Who's Who

• The Sales buddy is financially incentivized to close the deal and worries you might "screw the deal by being too honest." Halford warns against letting Sales make impossible promises — lying secures short-term deals but hurts everyone long-term.
• The champion (often a project manager) initiated the search, so their reputation is tied to success. Your job is to help them convince their decision-makers.
• When competing against a prospect's own engineering team (build-vs-buy), your approach doesn't really change — you still sell the value of your solution.

Get Into a Conversation

Halford believes limiting yourself to a scripted demo is "the biggest mistake you can make." His advice: turn the meeting into a conversation by asking open-ended questions that reveal pain points. Show only relevant features. Letting prospects talk makes them feel important — "people love to go on and moan about their work."

He suggests gentle name-dropping of similar clients and even underselling features: "it's a bit old. We're planning to revamp it soonish." This subtly signals high standards and trajectory. But he warns against coming off as a geek — "the prospect won't sign for something they don't understand."

Be Economical with the Truth

Halford acknowledges this is uncomfortable for technical people. In startups selling to enterprise, he argues, "faking it till you make it is necessary." The metaphor: tracks are laid while the train moves forward. If you trust your team to build what's needed, it's okay to be "a bit dishonest" about current capabilities. He explicitly says, "You won't be rewarded for always telling the truth."

He clarifies this isn't about malice — "I trusted our solution more than I did that of our competitors" — so overselling felt justified if it got a worried prospect over the line.

Phase 2: Pilots

The trickiest phase. The prospect is interested but not fully convinced. Halford's dating analogy: "this is the first time your date is staying at your place." All the veneer can vanish if you're sloppy.

He advises keeping pilots short and efficient. Longer pilots mean the prospect "will turn over rocks you don't want them to" and management may pull the plug entirely. Also, watch for Sales accepting pilots too freely — sometimes a thorough demo with customer data is sufficient.

Clarify the Success Criteria

Start with the prospect's specifications but don't accept them without review. Rephrase vague criteria into measurable, unambiguous ones. For example, instead of "Measure product environmental footprint," use "Measure GHG footprint for >95% of 2025 production volume." Splitting broad criteria into several specific ones makes them easier to tick off.

Even with rigid RFP processes, Halford encourages offering your own interpretation of criteria. The back-and-forth is worth it.

Control the Tempo

Weekly catch-ups are ideal. The pilot is "a race against time" — your objective is ticking criteria before the deadline. Prospects aren't tracking the timeline as closely as you are: "If you slip, they might not even notice." That's both a relief and a trap.

Always have an agenda. Without one, the prospect brings theirs, which may derail your priorities. "Controlling the tempo makes you look professional." Proactively surface issues in their data before they spot them — it builds trust.

Make Them Perceive Your Value

Don't just deliver results — narrate the difficulty. If integration took significant effort, explain why. "The last thing you want is to make your work appear trivial." But don't boast about implementation details they don't care about; focus on how the work simplifies their lives.

Space out what you show. A "steady drip of small wins is more memorable than a single firehose meeting." Remember: "the customer is buying you as much as the software."

Phase 3: Onboardings

The contract is signed (typically one-year minimum). This should be less stressful, but it tests the relationship — especially if Sales sweetened expectations during presales.

Don't Make It Drag On

Compress the onboarding as much as possible. Weekly meetings are ideal. The customer's patience wears thin the longer it takes. Same work in two months is better than four.

Be solution-oriented. Learn when to ask the customer to fix something versus fixing it yourself — an ability developed over time. If you ask someone to handle something and they don't follow through, "you're the one who will be blamed."

Halford advocates Extreme Ownership: "If something went wrong, it's on you." Blaming the customer is a dead end.

Build a Relationship

Treat the customer as a partner on equal footing, not a superior. "Make jokes. Onboardings can be stressful, but if no lives are at stake, then there's no reason for a tense atmosphere." Trust is earned through humility and efficiency.

He cites his company's motto: "Love creates growth." Well-treated customers become references who make or break future deals, especially for early-stage startups where prospects are skeptical.

Involve Your Product Team

Learn to say no to ad hoc work that doesn't fit the product roadmap. Log those demands into the product feedback pipe instead.

When something in the software is genuinely bad, deflect gracefully: "you're right, this part of the software is pretty bad, let me put you in touch with someone in our Product team." For impossible demands, let Account Management or leadership play "bad cop" while you remain the trusted good cop.

Parting Thoughts

Halford found being a solution engineer "a refreshing change from working on software in the background." It made him a better data scientist by building empathy for customers.

He believes many data scientists and engineers underestimate their value outside technical roles. "Plenty of companies still run on a swamp of ad hoc Excel files" — the bar is lower than it looks from inside a tech team. He notes that Forward Deployed Engineers and Product Engineers are increasingly sought after, and with AI commoditizing software, "your worth can increase by becoming a technical mind that can chat with customers."