Solution Selling: A Practical Discovery Process for Complex Deals

Solution selling is a consultative sales approach that diagnoses a buyer’s specific business problem before connecting an offering to the outcome they want. Use it when the problem, purchase, or implementation is complex enough that a product pitch alone cannot establish fit.

That means discovery is not a long prelude to a predetermined demo. The diagnosis may show that your offering is appropriate, that a smaller next step is better, or that the buyer should not purchase from you. The useful discipline is to separate what the buyer has confirmed from what the seller, a signal, or an AI system has inferred.

HubSpot describes solution selling as a consultative methodology that identifies specific pain points and presents a tailored solution. The discovery process, CRM field contract, and review controls in this guide are practitioner recommendations, not HubSpot defaults.

What is solution selling?

In a solution-selling conversation, the seller starts with the buyer’s situation rather than the product catalog. The seller asks how work is done today, where the problem appears, who is affected, what it costs, and what a successful change would mean. Only then does the seller decide whether and how to connect an offering to that need.

For example, “onboarding is slow” is a symptom, not yet a diagnosis. Discovery might reveal approval handoffs between operations and finance, inconsistent data, unclear ownership, or a seasonal staffing issue. Each cause implies a different response. The right outcome could be a configured solution, a process change, another investigation, or no sale.

HubSpot presents solution selling as particularly relevant to complex, customized, high-value, and multi-stakeholder sales. It also describes larger deals and stronger retention as potential benefits, not guaranteed results. Treat the method as a way to improve the quality of a buying decision, then measure commercial outcomes in your own context.

When is solution selling worth the extra time?

Choose the sales motion according to uncertainty and risk, not deal size alone. Consider these questions:

• Is the underlying problem ambiguous or still described only as a symptom?
• Will the offer require configuration, migration, integration, or change management?
• Could a wrong purchase create material financial, operational, or reputational cost?
• Are several teams affected or required to approve the change?
• Does the buyer need help developing the business case or implementation path?

When several answers are yes, deeper discovery is usually justified. Replacing a system across operations and finance, for example, may require agreement on the current process, migration risk, ownership, and success measures.

Use a faster product-led motion when the offer is standardized and low risk, the buyer already knows the problem and evaluation criteria, and one person can complete a self-serve or straightforward purchase. A lengthy qualification process can add friction without improving fit.

Solution selling vs. product selling

These are choices for different buying situations, not a claim that one method is always superior. In HubSpot’s framing, solution selling is more structured and problem-led, while consultative selling is broader. A consultative seller may understand the buyer thoroughly and still recommend no purchase.

Decision point	Solution-selling motion	Product-selling motion
Starting point	Buyer’s problem and desired outcome	Product, features, or specifications
Conversation	Diagnose, confirm, then recommend	Demonstrate, explain, then evaluate
Buyer experience	Explores impact, fit, stakeholders, and tradeoffs	Compares known specifications, availability, and price
Best fit	Complex, customized, or multi-team purchase	Clear need and standardized, low-risk offer

For a broader comparison of the terms, see HubSpot’s explanation of consultative selling.

A six-stage solution-selling discovery process

Good discovery produces a testable diagnosis, not merely a longer list of questions. Prepare a tentative theory, invite correction, listen for evidence, and use silence to let the buyer explain what matters. Record the buyer’s words and examples separately from your interpretation.

1. Prepare a testable hypothesis. Review relevant account context and write one tentative problem theory. Say, “I may be wrong, but it sounds like approval handoffs could be delaying onboarding. Is that relevant?” A hypothesis is a prompt for investigation, not a CRM fact.
2. Explore the stated problem. Ask the buyer to describe the current process and a recent example. Ask, “What happens when this process fails?” Capture the sequence, affected teams, frequency, and buyer-provided wording.
3. Establish evidence and impact. Ask, “How do you measure the impact?” Look for a buyer-provided measure, observed consequence, or named affected team. If no measure exists, record impact as unknown instead of inventing an ROI estimate.
4. Define the desired outcome and cost of inaction. Ask, “What would need to change for the outcome to count as successful?” and “What happens if this stays the same?” Keep the target and the consequence of inaction as separate fields.
5. Map decision authority and process. Ask who evaluates the tradeoffs, who approves the purchase, what criteria matter, and what steps remain. Record whether each detail is known, partial, or unknown. Do not treat a friendly contact as the decision-maker without confirmation.
6. Confirm the diagnosis before recommending. Summarize the problem, evidence, impact, desired outcome, stakeholders, and open questions. Ask the buyer to correct the summary. If impact or the decision path remains materially unclear, agree on the next discovery step rather than forcing a demo.

Frameworks can prompt useful questions, but none is required for solution selling. SPIN Selling organizes questions around situation, problem, implication, and need-payoff. MEDDIC provides a qualification structure that includes metrics and decision criteria. HubSpot’s source article describes PREDICT Selling as a framework used by its author’s organization. Use any framework to improve inquiry, not to replace the buyer’s answers.

Turn discovery into a usable CRM record

A CRM should preserve evidence and uncertainty, not just a polished seller summary. Define a small field contract before configuring a playbook or AI workflow:

• problem_statement: buyer-confirmed description, or a clearly labeled hypothesis.
• pain_evidence: concrete example, quotation, or observed consequence with its source.
• business_impact, desired_outcome, and cost_of_inaction: buyer-provided detail, or an explicit unknown.
• decision_maker_status and decision_process_status: confirmed authority and process, partial information, or unknown.
• timeline, source_type, source_url, source_record_id, captured_at, and captured_by: timing and provenance.
• review_status: for example, candidate, seller_confirmed, needs_follow_up, rejected, or incomplete.

These are proposed fields, not standard HubSpot properties. Keep AI-extracted candidates separate from seller-confirmed values. Require a human review state before an externally derived claim becomes part of outreach or a confirmed pain field.

HubSpot documents playbooks for contact, company, deal, ticket, and custom CRM records. Playbooks can capture notes, answers, and property-linked responses, and a submission can be logged as an engagement. The documentation lists applicable Sales Hub and Service Hub Professional and Enterprise conditions, assigned seat and permission requirements, and Enterprise-level conditions for direct property updates. Verify current subscriptions, seats, permissions, and configuration before rollout.

A practical sequence is to open the target CRM record, launch the configured playbook, capture the discovery answers, confirm the record association and critical values, and then log the submission. Property changes occur when the playbook is logged, not simply because a user typed an answer. If multiple questions target one property, review incomplete responses before logging so an unanswered question does not replace a useful existing value.

{
  "discovery_event_id": "illustrative-discovery-1042-2026-10-12T143000Z",
  "deal_id": "illustrative-deal-1042",
  "problem_statement": "Buyer reports approval handoffs delay onboarding",
  "pain_evidence": "Buyer described repeated operations-to-finance handoffs",
  "business_impact": "unknown",
  "desired_outcome": "unknown",
  "decision_process_status": "unknown",
  "source_type": "seller-entered discovery notes",
  "captured_at": "2026-10-12T14:30:00Z",
  "review_status": "needs_follow_up"
}

This is an illustrative record, not a HubSpot schema. The declared row grain is one completed discovery event. A separate signal, citation, AI run, or CRM property history entry needs its own identifier and record. Do not use a vague key such as company ID plus date when several events can occur on the same day.

For teams standardizing discovery data across systems, see ConsultEvo’s CRM systems and workflows service.

Where buyer signals and AI can help, and where they cannot

A company-level research or website signal can suggest a relevant discovery question. It does not establish an individual contact’s pain, authority, or readiness to buy. HubSpot documents buyer intent using tracking-code data, company matching, research topics, visitor activity, and configured criteria such as page paths. Its custom-signal documentation describes detected matches with summaries, source citations, and signal dates.

These features have account, permission, credit, tracking, refresh, and configuration conditions. Visitor-derived data may involve IP addresses or online identifiers, so privacy, consent, access, and retention controls belong in the implementation plan. Review the current buyer-intent documentation and custom-signal documentation before enabling them.

The following table separates the research trigger from the seller’s decision. The workflows are proposed operating patterns, not claims that HubSpot automatically performs every step.

Trigger	AI job	Validation	Action or fallback
Company-level intent signal	Summarize the signal and identify a possible discovery theme.	Check company match, signal date, source, freshness, and privacy basis.	Draft a question or research task. If the match is uncertain, ask the account owner to verify it.
Completed discovery notes	Extract explicit evidence, candidate themes, and unanswered questions.	Trace every claim to supplied notes; preserve unknown impact and authority.	Seller approves a call summary or follow-up task. Rejected extractions remain outside confirmed CRM fields.
Custom-signal citation	Organize the supplied summary and citations for review.	Inspect the citation, detection date, company association, and source relevance.	Store a candidate signal record. If the citation is stale or unsupported, mark it rejected rather than changing pain fields.

Use deterministic rules for blank required fields, stale signals, invalid associations, duplicate external IDs, ineligible deal stages, and missing consent. Use AI for note summaries, candidate themes, source organization, and draft follow-up questions. The difference matters because deterministic checks are easier to test and audit, while an AI interpretation still requires a seller’s review.

For a local signal dataset, define the row grain before building reports:

• One signal row represents one detected signal.
• One citation row represents one source attached to one signal.
• One AI run row represents one processing execution, including model and prompt version.
• One discovery-event row represents one completed call, meeting, or playbook submission.
• One aggregate row represents a defined reporting period and metric definition.

Use a vendor event or delivery ID when available. For a proposed local signal key, use company ID, signal type, detected timestamp, and a normalized source hash. For an AI run, include run ID, start time, engine, model, and prompt version. Enforce uniqueness with a database unique index or transactional upsert. A read-then-create check can race when concurrent workers process the same event.

For bounded AI workflow design and human review paths, see ConsultEvo’s AI agents service. For account-specific HubSpot configuration, see ConsultEvo’s HubSpot systems.

Measure process quality, not promised outcomes

Start with measures the team can inspect directly:

• Discovery completeness: the share of eligible opportunities with required fields completed, including explicit unknowns.
• Evidence-linked recommendations: the share of sampled recommendations tied to buyer-confirmed evidence and a stated desired outcome.
• Time to agreed next step: elapsed time from discovery to a buyer-agreed action for a defined cohort.
• Late-stage qualification gaps: how often material impact, stakeholder, or decision-process information is first found late in the cycle.

Define the eligible population, metric definition, and reporting period before comparison. Inspect a sample of records for evidence quality. Assess conversion, deal size, retention, or revenue separately, and do not treat a simple before-and-after change as proof that the discovery method caused a commercial outcome.

Solution-selling questions, answered

Is solution selling the same as consultative selling?

No. The approaches overlap, but HubSpot frames solution selling as a more structured process that diagnoses a specific problem before positioning an offering. Consultative selling is broader and can lead to a recommendation not to buy.

When should discovery stop?

Stop when the buyer has confirmed enough about the problem, impact, desired outcome, and decision path to test a recommendation. If a material gap remains, agree on the next discovery step instead.

Can solution selling work for low-cost or self-serve products?

It can, but a lengthy process may add friction when the product is standardized and the need is clear. Match discovery depth to purchase risk and uncertainty.

How do SPIN, MEDDIC, and PREDICT relate?

They are distinct frameworks that can structure parts of questioning or qualification. SPIN focuses on question types, MEDDIC on qualification, and PREDICT is described by HubSpot’s source article as a framework used by the author’s organization. None is a required standard for solution selling.

Make the recommendation earn its place

Solution selling is useful when diagnosis, stakeholder agreement, or implementation planning is part of establishing fit. It is unnecessary when the need and purchase criteria are already clear. Preserve buyer-confirmed evidence in the CRM, keep unknowns visible, and return to discovery whenever a material gap remains.