Evaluation Criteria & Scoring
Pick agencies like you hire a VP of growth. You need objective measures and gut checks. A simple rubric keeps politics out of the process.
Weighted scoring matrix
Score each vendor 1 to 5 across core dimensions, then multiply by weight.
- Strategic fit
- Domain expertise
- Technical/Martech capability
- Creative quality
- Execution reliability
- Measurement & analytics
- Commercial terms
Use a spreadsheet with rows for vendors and columns for each dimension. Add a final column for reference checks and a subjective “would I work with them” flag.
Suggested weightings by priority
If you want pipeline now:
- Execution reliability 20%
- Measurement & analytics 18%
- Demand gen capability 16%
- Creative 12%
- Martech 12%
- Domain expertise 10%
- Commercials 12%
If you want brand + long-term positioning:
- Domain expertise 20%
- Creative 18%
- Measurement 16%
- Martech 12%
- Execution 12%
- Demand gen 12%
- Commercials 10%
Adjust weights, but never put commercials above execution when you need results.
Minimum pass thresholds
Set non-negotiables. Any vendor that fails one of these gets cut.
- Execution reliability >= 3
- Measurement & analytics >= 3
- Martech compatibility >= 3
- At least two relevant case studies with verifiable outcomes
Capabilities checklist by function
Create a checklist you can tick during calls. Examples:
- Demand gen: SDR support, list build, outbound sequences, nurture workflows
- ABM: intent data, account tiering, personalized creatives, AB testing at account level
- Content/SEO: pillar content, topical authority plan, CMS migrations, structured data
- Paid media: audience modeling, multi-channel spend optimization, creative testing cadence
- Sales enablement: lead scoring matrix, SLA, playbooks, CRM integrations
- Analytics: event-level tracking, multi-touch attribution, dashboard templates, data exports
Red flags and deal-breakers
No written SLA on lead quality. No references you can contact. No access to creatives or source files. Overpromising with no experimental framework. Blurred ownership of data and reporting. If their “SEO” is just weekly blog posts with zero keyword research, walk away.
B2B Services & Deliverables
Stop accepting vague deliverables like “increase awareness.” Ask for artifacts. Here’s what real deliverables look like.
Demand generation tactics and outputs
Tactics must map to measurable outputs:
- Outbound: cleaned target lists, sequences, response scripts, conversion rate by stage
- Inbound: gated assets, landing pages, conversion optimization tests, MQL definitions
- Events: attendee lists, meeting outcomes, follow-up cadence
- Metrics: CPL, SQL rate, pipeline influenced value
Practical note: require sample outbound sequences and A/B test plans before signing. If they can’t sketch one fast, they won’t be iterative.
Account-based marketing components
ABM is a disciplined set of components, not a buzzword.
- Account selection framework with ICP scoring
- Personalized playbooks per tier (Tier 1 = bespoke, Tier 2 = programmatic personalization)
- Sales+marketing tasking sheet for each account
- Measurement: account engagement score, opportunities created by account tier
Mini scenario: for a 50-account Tier 1 program expect dedicated creative, a series of 3 touch orchestration, and a playbook to hand to sales for each account.
Content, SEO and thought leadership
Deliverables must be strategic and tactical:
- Topical map with keyword intent, prioritized by revenue potential
- Pillar pages, cluster articles, and gated playbooks mapped to funnel stages
- Thought leadership: interview transcripts, op-eds adapted for local press, executive-ready slides
- On-page and technical SEO task list with owner and timeline
Real test: ask for a 90-day content calendar showing promotion plan. If they only propose writing, they don’t get it.
Paid media and performance marketing
Outputs you should expect:
- Channel plan with hypothesis, audience, creative formats
- Creative variants and test matrix
- Weekly spend and performance report, with three optimizations recommended
- Attribution model they will use and how it will be implemented
If they pitch one channel for everything, you probably need to split work.
Sales enablement and lead handoff
This is where deals die. Demand clear mechanics:
- Lead scoring rules mapped to CRM fields
- SLA for response time by lead type
- Sales playbooks and email templates
- Handoff audit: sample leads moved through the funnel with timestamps
Insist on a dead-leads reprocessing workflow. If accounts accumulate without follow-up, the agency looks bad and you lose pipeline.
Analytics, attribution and reporting
Data deliverables should include:
- Event taxonomy and tracking plan
- Multi-touch model and rationale
- Daily/weekly KPI dashboard with raw exports
- Quarterly analytics review with experiment outcomes and recommended next steps
Red flag: proprietary black box attribution with no access to raw data.
Pricing, Contracts, Timelines
Money and legal details are where relationships fail. Be direct.
Pricing models explained (retainer, project, performance)
- Retainer: predictable month-to-month for ongoing programs. Good for steady-state demand gen.
- Project: fixed scope, defined deliverables. Use for migrations, audits, set-up work.
- Performance: payment tied to outcomes like leads or revenue. Only works with clean definitions and shared control over spend.
Don’t accept pure performance if you can’t control conversion steps. Agencies will underdeliver and blame your sales team.
Benchmark cost ranges and scope examples
Ballpark examples for Poland market realities:
- Small retainer: EUR 3k–6k per month. Basic inbound + light paid media.
- Mid-market: EUR 8k–20k per month. Multi-channel demand gen, content, ABM basics.
- Enterprise: EUR 25k+ per month. Full-funnel programs, dedicated team, analytics suite.
Project examples:
- CRM integration migration: EUR 8k–20k
- SEO migration + launch: EUR 10k–30k
Contract clauses to require or avoid
Require:
- Data ownership clause with export rights
- SLAs for lead delivery and reporting cadence
- Clear termination notice and handover obligations
Avoid:
- Long automatic renewal without performance review
- Exclusive lock-in on data or IP
- Vague success metrics
Pilot projects, ramp timelines, exit terms
Pilot design:
- 3 months minimum, 6 months ideal for meaningful signal
- Define sample size and success criteria up front
Ramp expectations:
- Month 1: set-up and foundational work
- Month 2: initial optimization and test runs
- Month 3 onward: stable delivery and scale
Exit terms: require a detailed handover package within 30 days of termination, including source files, tracking plan, and raw data exports.
Selection Process & RFP
Keep the process tight. Long RFPs attract boilerplate answers.
6-step vendor selection timeline
- Define outcomes and internal alignment
- Shortlist 4–6 vendors via referrals and case checks
- Send an RFP with a fixed response template
- Hold a 60-minute vetting call for shortlisted vendors
- Request a small paid pilot for final two
- Negotiate contract and start onboarding
Must-ask evaluation questions
- How would you run a 90-day test for our top ICP?
- What data will you need from our side on day one?
- Show me one failed experiment and what you learned.
- Who will actually do the work? Can I meet them?
- How do you calculate pipeline contribution?
RFP structure and required attachments
Keep it to one PDF plus attachments:
- One-page company brief and case studies with metrics
- Scope of work and list of deliverables
- Team bios and availability
- Price model options and assumptions
- Required attachments: sample playbooks, sample dashboard, two client references
Proposal scoring rubric template
Score 1 to 5 on:
- Understanding of brief
- Proposed approach and experimentation plan
- Team fit and availability
- References and outcomes
- Commercials and flexibility
Multiply by your predefined weights and rank.
Onboarding & Performance Governance
Poor onboarding wastes months. Set a strict 30–60–90 plan.
30–60–90 day onboarding checklist
30 days:
- Tracking, CRM, and access set-up
- Data and campaign baseline established
- Quick wins mapped
60 days:
- Initial campaign launches and A/B tests running
- First dashboard and weekly reporting cadence
- Sales enablement materials drafted
90 days:
- Optimization loop established
- First performance review with outcomes vs. KPIs
- Scaling recommendations
KPI dashboard and reporting cadence
Daily: system health metrics and spend burn
Weekly: campaign performance, tests, blockers
Monthly: pipeline formed, CPL, SQL rate, conversion ratios
Quarterly: strategic review, lift analysis, channel ROI
Provide raw data exports every month. No dashboards without CSV downloads.
Communication, governance and escalation paths
Set roles and a RACI chart. Insist on:
- Single point of contact on both sides
- Weekly tactical call, monthly strategy review
- Escalation path to senior leadership with SLA for responses
If vendor avoids naming senior contacts, assume they are outsourced.
Quarterly review and optimization ritual
Quarterly session should produce:
- What worked, what failed, why
- Two priority experiments for next quarter
- Budget reallocation plan with rationale
- Updated ICP or messaging tweaks based on evidence
Make decisions at the review. Meetings without decisions are theater.
Market Signals & Trends (Poland)
Poland is practical. Know the market quirks.
Martech compatibility and integrations
Poland adopters favor well-integrated stacks.
- Confirm CRM, marketing automation, analytics, and ad platforms all connect
- Ask about GDPR-compliant regional data handling and local CDNs for performance
- Expect translation of tracking tags and server-side tagging for cookie issues
Language, localization and compliance tips
Language matters more than you think.
- Use native copywriters for Polish. Direct translations feel off.
- Localize examples, case studies, and even pricing formats
- Compliance: screen cookie consent workflows and data processing addendums
Vertical specialization to prioritize
Certain verticals punch above their weight:
- SaaS targeting EU customers
- Manufacturing with export focus
- Fintech and regtech with localized compliance needs
Pick an agency that has a deep playbook for your vertical, not a mixed bag.
Emerging capabilities: AI and automation
Don’t buy hype.
- Expect automation for reporting, creative variants, and personalization tokens
- Real value: AI used for content ideation and segmentation, not replacing strategy
- Require transparency: any AI-generated content must be editable and traced to prompts
Final thought: treat agencies as partners who should reduce work, not add another layer of management. If they increase your meetings and produce fewer decisions, change them.