Vendor Decisions and Equity Impact

Let me tell you what happens when a library system signs a contract with a major discovery vendor without asking the equity question.

The vendor\'s system is fast, well-integrated, and - most importantly - cheap. Leadership looks at the contract and sees money saved. IT likes it because it integrates with the ILS. Circulation staff like it because it\'s user-friendly.

Nobody asks: Does this system work for blind patrons? Does it work for non-English speakers? What populations does this system actually serve well, and who gets left out?

By the time you realize the system doesn\'t screen-read properly for low-vision users, or that it only works in English, or that its recommendation algorithm actively filters out materials on topics that vulnerable communities search for, you\'ve already signed a three-year contract. Patrons who need accessible discovery can\'t use it. Immigrants who need Spanish-language browsing can\'t use it. Teenagers searching for LGBTQ+ resources can\'t find them because the algorithm deprioritizes "sensitive" topics. But you're locked in.

This is how vendor decisions shape equity. Not through intention - most vendors aren't actively trying to exclude vulnerable populations. Through inattention. Through procurement processes that optimize for cost and convenience instead of mission.

How Vendor Decisions Exclude: Five Real Mechanisms

1. Ebook Licensing Restrictions That Break Accessibility

You sign a major ebook licensing deal. The vendor\'s DRM (Digital Rights Management) system is strict - no downloading, no copying, no offline access. For the average reader on a desktop, it\'s fine. For a blind patron using a screen reader, it's unusable.

Screen readers need to access the underlying text. Many commercial ebook platforms use embedded DRM that prevents screen reader parsing. The patron sits in the library with access to thousands of ebooks, and can\'t read a single one because the vendor\'s licensing model doesn't support accessibility.

Low-vision patrons need to adjust fonts, colors, and contrast. Vendors often lock these controls behind paywall tiers or technical restrictions. The patron who needs text at 48pt can\'t get it because you licensed the "basic" tier. They have to choose: Use the library\'s ebook platform (and strain their vision) or buy from Amazon (where they can adjust accessibility settings).

This isn\'t negligence. Vendors understand accessibility law. They price accessible features as premium add-ons. Your procurement officer doesn\'t always notice the difference between a "standard" license and an "accessible" one - especially if the accessible version costs 30% more. So the library licenses the cheaper version. Blind patrons lose access. No one at the vendor's quarterly review calls this out.

2. Cost Structures That Force Hour Reductions in Low-Income Neighborhoods

This one I watched happen directly. A 15-branch system was paying each branch for its own discovery system license. When consolidating vendors, the new contract was cheaper per-branch - but only if all branches used the exact same platform. That sounds fine. It wasn't.

The unified platform saved $80K annually. But implementing it across all branches required IT retraining, migration costs, and 3 months of disruption. The CFO decided to fund the migration by reducing hours in the three lowest-circulation branches. All three were in neighborhoods where car ownership was low and patrons depended on the library for internet, job applications, and social services.

The decision logic: "These branches have lower circulation per dollar, so they\'re less efficient. Let\'s cut their hours to fund the transition." The actual impact: Patrons with the fewest resources got less access. An elderly immigrant who relied on daily library computer access for job searching now had 3 fewer hours per week to use the library. That\'s not a cost saving. That\'s a transfer of cost - from administration to vulnerable people.

This happens constantly because vendor contracts are evaluated on direct cost, not on downstream equity impact. A vendor deal that saves $100K but costs you equity in lower-income service areas isn\'t actually cheaper. You\'re just not measuring the real cost.

3. Data Privacy Gaps That Expose Vulnerable Populations

You sign with a vendor that uses patron search data for "analytics" and "algorithm improvement." The contract includes a clause about de-identifying data eventually, with no timeline. The vendor's privacy policy is carefully written to permit sharing with "service providers" without explicit patron consent.

For most patrons, this is annoying. For vulnerable populations, it\'s dangerous. An undocumented immigrant\'s search for "asylum information" is now tied to their account, sitting on a vendor\'s server. A domestic violence survivor\'s search for "protective order" is being tracked. A teenager's search for "transgender therapist near me" is data the vendor could legally share with marketing partners.

The vendor isn\'t acting maliciously. They\'re following their business model: patron data is valuable. But that business model treats vulnerable patrons" sensitive information as an asset to monetize. And it\'s legal because the contract language allows it, and the board never asked the privacy question during procurement.

4. Multilingual Service Gaps

Your community is 25% Spanish-speaking. You need a discovery system that supports Spanish interface, Spanish materials, Spanish search. You get quotes from three major vendors. All of them offer "Spanish translation" as an add-on. Only one actually has Spanish-language indexing and search optimization built into the platform.

The vendor with real Spanish support costs 40% more. The director asks: Can\'t we just translate the interface? Yes, your IT department can do a quick machine translation. It will work okay for menus. It won\'t work for search logic, for materials discovery, for results ranking. The Spanish-speaking patron will get inferior search results - fewer hits, less relevant materials, poorer discoverability - because the vendor's system was designed for English first, Spanish never.

You\'ll probably license the cheaper version. The Spanish-speaking community gets access to a system not built for them. They\'re technically welcome to use it, but in practice, they get worse results. That\'s not equal access. That\'s the illusion of access.

5. Algorithmic Bias That Filters Vulnerable Populations Out of Discovery

Your discovery system uses machine learning to rank results. It learns from aggregated patron behavior: what people click on, what they check out, what they rate. Sounds neutral. It's not.

The algorithm learns that materials on certain topics (LGBTQ+ health, immigration law, domestic violence, racial justice) are clicked on less frequently by the general population. So it deprioritizes them in results - they show up on page 3 instead of page 1. The algorithm is trained on the behavior of people with privilege (who don't need to search for these topics) and learns to bury resources for people who do.

A teenager searching "gender dysphoria" gets a general Wikipedia definition on page 1 and library materials on pages 3-5. That\'s not because your library doesn\'t have good materials. It\'s because the algorithm doesn\'t think they're relevant. The teenager gives up and searches Google instead, where they find less trustworthy sources.

This bias isn\'t malicious. It\'s the natural outcome of training algorithms on majority behavior and calling it "relevance." But it systematically deprioritizes vulnerable populations" needs. And you won\'t know it's happening unless you specifically test for it during vendor evaluation.

Case Study: The Branch Hour Reduction (Real Story)

Here's what actually happened in a 12-branch system I consulted with:

The Setup: Three branches were in lower-income neighborhoods. They had lower circulation than branches in wealthier areas. When the annual budget came up, administration proposed consolidating to a single discovery vendor (cheaper) and cutting hours at low-performing branches to fund the transition.

The Decision Logic: "The downtown branch circulates 45,000 items/year on 60 hours open per week. The East End branch circulates 12,000 items/year on 60 hours open per week. The downtown branch is 3.75x more efficient. We should shift resources there."

What Nobody Asked: Why is circulation lower at the East End branch? Is it because the community doesn\'t need library services, or because the community\'s needs are different?

The real answer: The East End branch served 80% non-English speakers (predominantly Spanish and Vietnamese), 40% unhoused, 35% disabled, and 25% undocumented. The high school branch downtown served affluent teenagers who used the library primarily for studying and socializing. Its circulation was high because teenagers check out multiple books per visit.

But the East End branch\'s impact was different. It was where immigrants came to access job application help, where unhoused patrons could sit safely, where people navigated benefits applications. It had lower checkout counts because the community\'s primary need wasn't books - it was access, assistance, and safety.

When hours were cut from 60 to 48 per week, what happened? Job seekers lost hours they needed to complete applications. Unhoused patrons lost access to services provided only during certain hours. Immigrant families with 9-to-5 jobs lost evening hours.

The downtown branch kept 60+ hours and expanded weekend access. Total library system efficiency went up on the spreadsheet. The library's equity went down in reality.

Equity Assessment Framework for Vendor Decisions

Before signing ANY vendor contract, conduct this assessment. Don't skip it. This is what separates a library that serves its mission from a library that serves its budget.

Step 1: Identify Vulnerable Populations You Serve

Who are the populations most dependent on your library? List them specifically:

• Blind and low-vision patrons (what % of your community?)
• Deaf/hard of hearing patrons
• Patrons with mobility disabilities
• Non-English speakers (which languages?)
• Undocumented immigrants
• Unhoused patrons
• Domestic violence survivors
• LGBTQ+ youth
• People with cognitive disabilities
• Lower-income patrons dependent on free public access

Get actual numbers where possible. "We serve some Spanish speakers" is not a useful assessment. "25% of registered patrons have Spanish listed as primary language, with 40% of those not speaking English fluently" is useful.

Step 2: Test the Vendor Against Each Population's Needs

For each vulnerable population, ask:

Don\'t accept vendor assurances. Actually test with real patrons from vulnerable populations. Bring a blind patron in with their screen reader. Bring a Spanish speaker and do search tests. This takes 2-3 hours. It\'s the most important hours you'll spend on procurement.

Step 3: Quantify Equity Impact

For each vulnerable population, estimate:

• Access Impact: How many patrons will this exclude or inconvenience? (5 blind patrons? 200 Spanish speakers?)
• Severity: Is this a convenience issue or a critical barrier? (Can't enlarge fonts = critical; interface slightly slow = convenience)
• Alternatives: Can patrons use other library services if this vendor fails them? (If a blind patron can't use discovery, can they call reference? Can reference staff help?)
• Cost to Fix: What would it take to make this accessible? (Add-on module? Staff workaround? Different vendor?)

Step 4: Make a Decision

You have four options:

Option A: Don\'t Use This Vendor If the equity cost is high and the vendor won\'t fix it, don\'t sign. Find a vendor that actually serves your whole community. This is a legitimate business decision, not a failure. You\'re saying: "Our mission includes all patrons. We won't compromise that for cost savings."

Option B: Demand Contract Changes Tell the vendor exactly what needs to change: "We need Spanish language support built into the platform (not translated), accessibility certified for WCAG 2.1 AA, and commitment to bias testing. If you can\'t do it, we\'ll choose another vendor." Many vendors will make changes if you represent enough potential business. Smaller vendors especially - they're often more flexible than massive corporations.

Option C: Implement with Compensatory Services If the vendor has gaps, fill them with staff. "The vendor\'s discovery system isn\'t accessible for blind patrons, but our reference staff can provide personalized search assistance." (Note: This is more expensive than just hiring an accessible vendor, but it's honest about the tradeoff. Many libraries choose this.)

Option D: Implement with Equity Monitoring You\'re signing despite known gaps. Document them. Plan to reassess in 6 months. If patrons from vulnerable populations aren\'t using the vendor, commit to changing. Don't just accept mediocre equity outcomes.

Whatever you choose, write down your reasoning. "We chose Vendor X despite lack of Spanish language support because Option A (finding accessible vendor) took 6 additional months and we needed to implement by fiscal year. We're providing Spanish-language reference support as a compensatory measure and reassessing in 12 months." This creates accountability. Future directors can see why this choice was made and whether it was worth it.

Vendor Equity Audit Checklist

Use this before every contract renewal or new vendor evaluation:

Board-Level Equity Accountability for Vendors

Here's what boards should ask about vendors:

At Contract Renewal Time

• Is the vendor accessible to blind/low-vision patrons? How do we know? (Cite specific tests.)
• Does the vendor support languages our community speaks? What gaps exist?
• What patron populations are using this vendor vs. avoiding it? Are the disparities explained?
• Have we monitored vendor impact on service hours in lower-income areas?
• Does our contract include strict data privacy protections for vulnerable populations?
• What's happened in the past year if this vendor had equity issues? Were they fixed?

When Implementing New Vendor

• How did we test this vendor with vulnerable populations?
• What equity gaps does this vendor have, and how are we addressing them?
• Will implementation require service cuts in lower-income areas?
• What's our equity monitoring plan for the first 12 months?
• If equity outcomes are worse than projected, what will we do? (Is there an exit clause?)

In Annual Performance Review

• Usage statistics broken down by patron demographics - who's using this vendor?
• Complaints or accessibility issues reported by vulnerable patrons?
• Any security/privacy incidents? How were they handled?
• Actual cost to equity (hours cuts? populations excluded?) vs. projected cost?

Contract Language: Equity Clauses

Add these to every vendor contract:

Why Boards Need to Care (Beyond Mission Alignment)

Risk Management: An inaccessible vendor exposes you to ADA liability. A vendor with poor data security exposes vulnerable patrons to harm and you to litigation. Vendor decisions are risk management decisions.

Public Trust: When word gets out that your library uses an inaccessible discovery system, or that you cut hours in the neighborhood with the most vulnerable populations, communities stop trusting you. Trust is your most valuable asset. Vendor decisions affect it.

Mission Alignment: Libraries exist to serve all community members, especially those with the fewest resources. Every vendor decision either advances or undermines that mission. If your board approves a vendor knowing it excludes vulnerable populations, and you don\'t fix it, you're not serving your mission. You're just claiming to.

Equity as Fiduciary Duty: Boards have fiduciary duties to the public they serve. That includes equity obligations. A procurement decision that saves $100K but harms vulnerable populations isn\'t a good financial decision - it\'s a breach of duty to the communities you serve most.

Related Reading

Deepen your understanding of vendor relationships, privacy, and equity in library technology:

• Privacy Policy AI Fixes: What to Demand — How to update your privacy policy to address AI vendor practices and protect patron data from surveillance.
• Colorado AI Act: What Libraries Need to Know — Emerging AI regulation that affects vendor contracts and procurement decisions.
• Surveillance or Service? AI Privacy for Vulnerable Patrons — How AI systems disproportionately harm at-risk community members and what vendors promise (and deliver).