Reference bank: managing and formatting references for bidders

References serve two distinct purposes in public procurement, and bidders must understand the difference. The first is a mandatory selection criterion: 'The bidder must have at least 3 comparable deliveries in the past 3 years.' If the requirement is not met, the tender is automatically rejected — no points are awarded and there is no discretion.

The second purpose is as a scored evaluation criterion: 'Relevance of references: 0–10 points.' Here, better references yield more points and can decide the tender. How references are formatted and selected is critical — the same references can produce very different scores depending on how they are described and which ones are chosen.

The contracting authority is looking for an answer to one fundamental question: can this bidder deliver for us? A reference answers this by demonstrating that the bidder has completed a comparable delivery before, the scope and complexity are of the same caliber, the client was satisfied, and the delivery was on schedule and within budget.

References are also a strategic weapon: well-chosen and well-described references can produce a significant point advantage in quality evaluation. This means the time invested in managing and formatting references pays for itself many times over. A systematic reference bank is therefore every bidder's essential tool.

The foundation of a systematic reference bank is a standardized reference card that collects the same information for every delivery. Basic information includes project name, client (organization), sector (public/private), industry (healthcare, IT, construction, etc.), contract period (start date–end date), total contract value in euros, and annual value in euros.

Content information covers the description (2–3 sentences on what was done), scope (number of users, volume, geographic area), technologies and methods used, and special features (what made the delivery demanding or distinctive). This information serves as raw material for reference descriptions, which are tailored for each specific tender.

Contact information includes the reference person's name, role/title, phone number, and email address. Additionally, record the client's satisfaction rating (excellent/good/satisfactory), keywords for search purposes, and the last update date. The reference person's contact details are critical information that becomes outdated quickly.

Collect reference information immediately after project completion — details are fresh in memory and the reference person is reachable. Request feedback and contact details from the client, document scope, value, and results concretely, and agree with the reference person that you may use them as a reference in future tenders.

An effective reference bank is organized along multiple axes so the right references can be found quickly for each tender. Industry classification (healthcare, IT, real estate, consulting, construction, education) is the most obvious method but insufficient on its own.

Service type classification differentiates projects, ongoing services, consulting, product deliveries, and maintenance. This is important because contracting authorities often specifically seek experience with a corresponding service type — project experience may not convince in a continuous service procurement.

Client type classification divides references into deliveries for government, municipalities, hospital districts, universities, public enterprises, and private companies. Public sector references demonstrate familiarity with the public sector operating environment, which is often valuable. However, private sector references are generally accepted — the Procurement Act does not restrict references to public sector deliveries.

Size classification divides references into three groups: small (under EUR 50,000), medium (EUR 50,000–500,000), and large (over EUR 500,000). This is essential because many tenders set a minimum value for references. Keep these classifications up to date and add new references to the bank as soon as they arise.

When a request for proposals arrives, reference selection proceeds in three phases. In the first phase, ineligible references are eliminated: references that are too old (typically a 3–5 year time limit), too small (below minimum value), and references from the wrong industry or service type. After this screening, the remaining references are potential candidates.

In the second phase, remaining references are scored using a weighted matrix. Relevance to the procurement object receives the highest weight (30%), as it is the contracting authority's most important evaluation criterion. Scope and value correspondence is weighted at 20%, client type match (public/private) at 15%, reference person's reachability at 15%, recency at 10%, and special feature relevance at 10%. Each criterion is rated on a scale of 1–5.

In the third phase, the highest-scoring references are selected. If the RFP requires 3 references, select the top 3 — but also ensure diversity. Different clients and different perspectives on the service are more convincing than three very similar references from the same client.

The selection strategy is not a one-time exercise but a continuous process. References are re-selected for every tender based on that specific procurement's requirements. The same reference may be the top choice in one tender and be eliminated in another. This is why the reference bank must be sufficiently broad and diverse.

When the RFP does not prescribe a format, a reference description is built from five parts. The header (1 line) contains the client, project name, contract period, and value. The procurement description (2–3 sentences) explains what the client needed and why — this contextualizes the reference for the evaluator.

The delivered solution is the core of the description (3–5 sentences): what was concretely done, what was the scope (number of users, volumes, geographic coverage), and what technologies, methods, or standards were used. In this section, the bidder demonstrates the depth of their expertise and the complexity of the delivery.

Results (2–3 sentences) are the most convincing part: measurable outcomes such as 'service level 99.8%' or 'delivery 2 weeks ahead of schedule' are many times more effective than generic descriptions. Client feedback or a contract extension is a strong signal of satisfaction.

The final section is relevance to this procurement (1–2 sentences), where the reference is linked directly to the specific RFP. This is the section that is tailored for each tender individually: 'This delivery corresponds to the current procurement particularly regarding X and Y requirements.' Tailoring demonstrates that the bidder is not using a generic copy.

Use numbers whenever possible. 'We served 15,000 end users across 12 locations' is significantly more convincing than 'we served a large user base across multiple locations.' Numbers are concrete, comparable, and memorable. Euro amounts, user counts, percentage improvements, and timelines are the most effective language.

Highlight relevance to the RFP requirements explicitly. If the RFP requires cloud migration experience and your reference includes cloud migration, bring it up directly: 'The project included a 500-user cloud migration to an AWS environment.' The evaluator should not have to guess at relevance — make it easy for them.

Be precise in describing the execution. 'We delivered a comprehensive service including design, implementation, testing, and deployment' says significantly more than 'we handled the project from start to finish.' Break the whole into parts so the evaluator can see everything that was included in the delivery.

Avoid corporate jargon and industry-specific terminology. The contracting authority's evaluator may not be familiar with your industry — write so that even a non-expert can understand. This does not mean oversimplifying but rather clarity: technical content can be described precisely without abbreviations and professional slang.

Managing reference contacts is the reference bank's most critical maintenance task. People change jobs, contact details become outdated, and a reference person may be unreachable at a critical moment. Update contact details at least once a year and keep 2 contacts for each reference (primary and backup).

Always notify the reference person in advance when using them as a reference — a surprise call from a contracting authority is not a good experience for anyone. Request a LinkedIn connection as a permanent contact so you can find the person even if they change organizations. Remember that a reference person's positive and informed response to a contracting authority's verification call can decide the tender.

References expire: most RFPs require references from the last 3 or 5 years. This means every year some references drop out of the usable pool. You need new references continuously, and collecting them is a strategic decision — choose projects that will produce good references for future tenders.

Practical maintenance routine: set a calendar reminder for a semi-annual reference bank review. Check contact details, remove expired references from active use (but retain in archives), add new deliveries, and assess whether there are gaps in your reference portfolio relative to upcoming tenders.

If you don't have enough of your own references, the first option is a consortium. A partner's references complement your own, and consortium members' references are counted together. This is a particularly effective strategy for new companies or companies expanding into a new industry.

The second option is relying on subcontractor capacity (ESPD Part II C). A subcontractor's references can be used if the bidder relies on the subcontractor's resources to meet selection criteria. In this case, the subcontractor fills out their own ESPD and commits to the delivery. This is legally stronger than a simple reference list.

The third option is personnel references. In some tenders, key personnel's experience is evaluated instead of the company's. A new company with experienced specialists can meet requirements through individuals' references. Check the RFP for whether personnel references are accepted.

Confidential references are their own special situation. If a reference is subject to a non-disclosure agreement, you can still report it at a general level: 'A large Finnish healthcare organization, contract value approximately EUR 500,000.' The reference person can confirm the details directly to the contracting authority. Explain in the tender why more detailed information cannot be provided — this is standard practice and does not diminish the reference's value.

Private sector references are generally accepted in public procurement — the Procurement Act does not restrict references to public sector deliveries. However, public sector references are a 'bonus' because they demonstrate familiarity with the public sector operating environment. In private sector references, clearly describe the relevance to public procurement.

Quality over quantity: if the RFP requires at least 3 references, submit exactly 3 of your best — no more. Extra references do not earn additional points and may even dilute the impression. Exception: if the evaluation criteria award points for additional references.

The time frame is critical: a 3-year time limit generally refers to the contract's end date, not its start date. This means a long-term contract that ended 2 years ago is typically eligible. Check the RFP for the exact time limit definition — interpretations vary.

A reference bank is a competitive advantage that takes time to build but pays dividends with every tender. Start systematic collection today: document ongoing deliveries, request feedback from clients, and build a reference card for every significant project.