Database Administrator Job Description Template 2026

Last updated: May 16, 2026

A database administrator owns the production data tier, tunes the queries that decide whether the app feels fast, and runs the backup-and-recovery posture the business actually depends on, with 2026 U.S. base salaries from $110,000 at mid-level to $215,000 for senior cloud platform DBAs. Two thirds of the DBA postings we benchmark on intake calls are still written for a role that no longer exists, the all-in-one DBA who patches the OS, runs the backup tape job, and tunes the optimizer all from the same chair. The template below is the shape that closes inside 30 days when hiring managers commit to the actual 2026 archetype and post the comp band.

Our IT desk gets a steady drip of DBA intake calls that all start the same. A senior DBA who has been at the company for nine years just gave notice. The replacement requisition reads exactly like the JD that hired him in 2017. Nobody on the team can remember whether the role still needs hands-on Oracle RAC experience or whether the workload moved to managed Postgres on RDS two years ago. By the time the JD reaches our desk, it is a Frankenstein of every database the company has ever touched and a list of certifications that have been deprecated for five years.

The recruiting problem is downstream of the spec, and on our database administration staffing desk the fix almost always lives in the JD, not the candidate pool. Senior DBAs in 2026 are mostly platform engineers who happen to specialize in the data layer. They live in Terraform, they argue about query plans on Slack, they have opinions about replication topology, and they are not interested in a role that reads like a 2015 night-shift on-call rotation. Posting the old JD pulls candidates from the bottom of the market who do not match what the team actually needs, and the strong candidates the team wants scroll past without applying.

Tom Kenaley here. Senior Partner and President at KORE1. The engineers on our IT desk are the ones tuning queries and parsing execution plans on candidate calls. My side of the house sits closer to the requisition shape: which clients keep refilling the same role, which JDs land hires quickly, which ones come back to us twice. KORE1 collects a placement fee on the hire, so weight the recommendation accordingly. The JD template below works the same whether you partner with a staffing firm or post the role yourself.

Three Archetypes the Title Hides in 2026

Database administrator in 2026 collapses three distinct roles into one title: production DBA, cloud platform DBA, and data engineer with DBA responsibilities. Each lives on a different team, screens for a different skill set, and lands in a different comp band roughly $25,000 to $40,000 apart.

Picking the archetype is not optional. It is the difference between a 30-day search and a 90-day one with two intermediate JD rewrites.

Production DBA. The shape most hiring managers still picture when they say “we need a DBA.” Owns the live database tier for one or more business-critical systems. Patches, parameter tunes, sizes the buffer cache, runs the failover drills, owns the runbook for a corrupt redo log at 2 a.m. Lives close to the underlying platform, whether that is on-premises Oracle, SQL Server on Azure VMs, or Postgres running on EKS. Mid-level production DBAs in 2026 run $105,000 to $135,000 across the metros where our IT desk does most of its placement volume. Senior production DBAs land at $135,000 to $180,000 in non-coastal metros and $160,000 to $210,000 on the coasts, with the band widening at companies that run mission-critical Oracle or have a real on-call burden. Principal DBAs at financial services, healthcare IT, and large e-commerce push higher. The credentials that still matter at this archetype are Oracle Certified Professional, Microsoft Certified: Azure Database Administrator Associate, and increasingly the Postgres-specific EnterpriseDB credentials. Hands-on experience recovering from a real incident is the better signal, but the cert filters the resume pile.

Cloud Platform DBA. Different team, different skill set. Sits next to the platform engineering function or inside it. The work is half database, half infrastructure-as-code. Builds the Terraform modules that provision RDS, Cloud SQL, or Azure SQL fleets, designs the cross-region replication, owns the cost model, and runs the migration projects when the company finally retires the last on-prem Oracle box. The candidate pool is smaller than for production DBA because most production DBAs never learned Terraform deeply, and most platform engineers never owned a query plan. Comp lands $130,000 to $200,000 for mid through senior. The roles that close fast post the cloud provider, the database engine, and the migration runway up front. The roles that drag say “cloud experience preferred” and end up open for six months.

Data Engineer with DBA Responsibilities. The hybrid the analytics-driven companies actually want. Sits on the data platform team. Builds and owns the warehouse on Snowflake, BigQuery, or Databricks. Writes the dbt models, owns the ELT pipelines, partitions the fact tables, and gets pulled into transactional database work when a service team needs a query plan reviewed or a partitioning strategy critiqued. Comp pulls $125,000 to $190,000 for mid through senior, higher at companies where the data platform is the product. Misposting this as a generic “DBA” pulls candidates who have never seen a window function. Misposting it as a “data engineer” pulls candidates who can build a Snowflake table but go silent when asked about index design on a 200-million-row OLTP system.

Pick the archetype. Write the JD for that shape. The version of this role that closes inside our 30-day median is the one where the hiring manager can answer, in a single sentence, which engine the DBA will own and which team they will sit on in the first 90 days.

Sources: ZipRecruiter (April 2026), Glassdoor (2026), Levels.fyi (2026), KORE1 internal placement data 2025-2026. 25th to 75th percentile. Coastal metros (SF, NYC, Seattle, Boston) add 15-22%. Fintech, healthcare IT, and ad-tech add 10-18%.

The DBA Job Description Template

Copy what fits. Cut what does not. The brackets are placeholders for your actual engine, your actual hosting, your actual on-call rotation. The parentheticals are notes to the person writing the post and should never appear on the live listing.

Job Title

[Database Administrator (Production) / Database Administrator (Cloud Platform) / Data Engineer, Database]

(The qualifier in parentheses is the most useful word in the title. Skip it and the wrong half of the candidate pool applies. Add it and the JD does half the screening for you. Avoid “Sr. DBA / Data Engineer” with the slash. It looks like the hiring manager could not decide.)

About the Role

(Three sentences. What database does this person own? Who do they sit with? Is the role remote, hybrid, or onsite? Skip the company mission paragraph. Resume readers scroll past it.)

[Company Name] is hiring a [DBA archetype] to own [specific scope: our Oracle 19c production tier supporting the order management system / our 40-cluster RDS Postgres fleet running 200 microservices / our Snowflake warehouse processing 2 billion daily events]. You will partner with [platform engineering / data engineering / a product squad] and report to [Director of Infrastructure / VP Data / Head of Platform]. The role is [remote in the U.S. / hybrid in {city} / onsite in {city}] with [primary / secondary / shared] participation in a [follow-the-sun / U.S.-only / business hours only] on-call rotation.

What You Will Own

(Six specific responsibilities. Each one should describe something this DBA would do in the first 90 days. Avoid the generic “ensure database availability” line. Name the engines, the scale, the migration projects you want them to inherit.)

• Own the production posture of [specific database systems], including the runbook coverage, the backup-and-recovery posture validated by quarterly restore drills, and the on-call response that the rest of engineering depends on
• Lead query and schema performance work for [scope], partnering with application teams on index design, partitioning strategy, and the slow-query reviews that prevent the next late-night incident
• Manage upgrade, patching, and version-lifecycle planning for [Oracle 19c-23c / SQL Server 2022 / PostgreSQL 15-17 / MySQL 8 / specific engines], including the testing rigor that catches regressions before they reach prod
• Build and maintain infrastructure-as-code for the database tier in [Terraform / Pulumi / CloudFormation], with module design and state isolation that other platform teams can reuse
• Own the data protection model end-to-end, including encryption posture, access controls, audit trail, and the compliance scope that your auditors actually care about ([SOC 2 / HIPAA / PCI / GDPR])
• Partner with [security / data platform / application engineering] on capacity planning, cost optimization, and the migration projects that will retire [specific legacy system] over the next [number] quarters

What You Bring

(Required and preferred, split clearly. Pad the required list and the pool shrinks. The candidates who would have been excellent self-select out the moment they fail a bullet that did not actually need to be there.)

Required:

• [5-7 / 7-10 / 10+] years of production database administration experience with at least [number] years on [specific engine you actually run]
• Deep hands-on experience with [specific engine and version], including upgrade work, backup-and-recovery design, and at least one incident you helped recover from that you can describe in interview
• Performance tuning fluency, including index design, query plan reading, statistics management, and the optimizer quirks of your stated engine that only show up in production at scale
• Strong scripting and automation skills in [Python / Bash / PowerShell], evidenced by tooling you wrote and can describe, not just inline shell glue
• Production experience with [Terraform / CloudFormation / Pulumi] for database infrastructure provisioning, including state management and pipeline integration
• Comfort with on-call. Specifically, comfort with a [primary / rotating] page schedule that includes [overnight / weekend] coverage for production data incidents

Preferred:

• Oracle Certified Professional, Microsoft Certified: Azure Database Administrator Associate, AWS Certified Database Specialty, or EnterpriseDB Postgres Professional
• Experience with [Snowflake / Databricks / BigQuery] at production scale, including warehouse design, role-based access, and the cost model conversation with finance
• Experience with database migration work, especially heterogeneous migrations such as Oracle to Postgres or SQL Server to Aurora
• Replication topology experience: streaming replication, logical replication, multi-region failover, and the data-loss windows each pattern actually buys you
• Vertical experience in [your industry: fintech, healthcare IT, ad-tech, e-commerce, gaming]
• Background in data security, including row-level security, column-level encryption, and the compliance regime that applies to your data class

Compensation and Benefits

(Post the range. Twelve states and a growing list of metro ordinances now require it, the cost of leaving it off has been measured by every aggregator that publishes time-to-fill data, and postings without ranges fill slower across our IT desk by roughly two weeks. A $35K band is fine and signals seniority without locking you into a number.)

• Base salary: $[X] – $[Y]
• [Additional components: equity grant, signing bonus, on-call differential, performance bonus]
• [Three or four benefits that actually matter: healthcare details, 401k match, PTO model, remote stipend, learning and development budget]

About [Company Name]

(Two or three sentences. Product, stage, team size, why a senior DBA should care. Skip the boilerplate that begins with “We are a fast-growing.”)

Where Most DBA JDs Lose the Right Candidates

Four mistakes show up in maybe four out of every five DBA postings we benchmark on intake calls. None are subtle. All are fixable in one editing pass.

Listing every database engine. “Expert in Oracle, SQL Server, MySQL, Postgres, MongoDB, and Cassandra.” That candidate is a vendor sales engineer, not a working DBA. Real DBAs go deep on one engine and have working knowledge of one or two others. The very small population that has actually operated all six engines at production scale are either consulting independently at $250 an hour, running platform engineering at a hyperscaler, or no longer on the market at any number you have approval to pay. Six-engine bullets signal a JD that nobody on the engineering bench reviewed. Name the engine that holds production data. Add a second only if a real migration is on the roadmap that the new hire will own.

The on-call line buried in the benefits paragraph. If the DBA will carry a primary pager every third week, say so in the second paragraph. Burying that detail until the offer stage costs hires. One Phoenix client opened a Postgres DBA search in November, ran four candidates through five interview rounds each, and watched the top two pull out after the final-stage call surfaced a rotation they had not been told about. Three of the four were strong fits we had pre-screened against the spec the client first gave us. The eventual fill landed at week 14. The next role the client opened with us, a SQL Server platform DBA, named the rotation cadence in the second paragraph and closed in 41 days.

Asking for cert lists that have been deprecated. Oracle 12c OCP, MCSA SQL Server 2016, MongoDB Certified DBA Associate. Three credentials we still see required on live postings in 2026. Oracle 12c reached end-of-life in 2022. The MCSA path was retired by Microsoft in 2021. The MongoDB DBA Associate has been replaced twice. Required-cert lists copy-pasted from a 2018 JD signal to senior candidates that the hiring manager does not know what is current, and the candidates who would otherwise apply assume the rest of the spec is similarly stale. Audit the cert list against the vendor’s current credential page before posting.

Generic responsibilities that read like a software brochure. “Ensure high availability.” “Drive operational excellence.” “Champion data best practices.” Senior DBAs scroll past those bullets. They want specifics. What engine? What version? What is the alert volume on a normal week? How big is the largest table? Are we running on bare metal, on VMware, on EKS, on RDS, on Azure SQL Managed Instance? Name the scale. Name the version. Name the migration the new hire will inherit. The candidates worth hiring will read those specifics and self-screen with more accuracy than your recruiter ever could.

Engines and Platforms to Name By Name

Generic DBA skill lists pull generic resumes. Name the engines you actually run. The candidates you want already match the production environment they are leaving. The ones who do not match self-select out, which is the goal.

Relational engines. PostgreSQL has overtaken Oracle as the most-listed engine in DBA postings on our IT desk over the past 18 months, but Oracle still dominates the high end at financial services and the regulated verticals where the migration has not happened yet. SQL Server holds Microsoft-shop majority. MySQL persists at companies with deep MySQL Aurora investments. Naming “relational database experience” without naming the engine reads as a hiring manager who has not asked their senior engineers what they actually run.

Managed cloud database services. Amazon RDS and Aurora dominate AWS-heavy shops. Google Cloud SQL and AlloyDB show up in GCP-native organizations. Azure SQL Database and Azure Database for Postgres own the Microsoft estate. Snowflake and Databricks are warehouses, not OLTP engines, and listing them under “DBA experience” without distinguishing transactional from analytical work is a signal that the JD was written by someone in HR. Name the service. Name the version. Name whether you are running self-managed or managed.

NoSQL and specialized stores. MongoDB, Cassandra, DynamoDB, Redis, Elasticsearch. Each one screens for a different operational skill set. Listing all five as “preferred” pulls candidates who used one of them in a hackathon. Pick the one that holds production data and ask for real experience with it. The other four can be optional.

Infrastructure-as-code and automation. Terraform is the dominant choice for database provisioning in 2026. Pulumi has meaningful adoption. CloudFormation appears in AWS-native shops that resisted Terraform. Ansible still owns post-provisioning configuration for on-prem and VM-based deployments. Listing “infrastructure-as-code experience” without naming the tool reads as a JD nobody on the engineering bench reviewed.

Observability and monitoring. Prometheus plus PostgreSQL exporter or pg_stat_statements for self-hosted Postgres. Datadog Database Monitoring for managed-SaaS shops. SolarWinds DPA, Quest Foglight, and Redgate persist in SQL Server estates. Oracle Enterprise Manager still anchors most large Oracle installations. Naming the monitoring stack is screener-level information that the strongest DBAs use to decide whether to apply.

DBA Salary Benchmarks 2026

Four sources, real spread. The aggregator data underweights the high end because cloud platform DBA roles at infrastructure-heavy companies do not always end up in public surveys under the “database administrator” title. KORE1’s placement data sits at the senior end because that is where most of our IT desk volume lives. The Bureau of Labor Statistics groups DBA work under a broader Database and Network Administrators category, which dilutes the median because it lumps in adjacent network and systems roles.

The Bureau of Labor Statistics projects 8% growth for database administrator roles through 2033, with about 8,800 annual openings. The headline number understates the real demand shift. Production DBA job count is roughly flat. Cloud platform DBA and data-engineer-with-DBA-skills count is growing fast enough that the BLS category lumps them with adjacent roles that have already absorbed the headcount. Senior cloud DBAs and data-platform DBAs are the genuinely short pool in 2026.

Metro premium math is the next thing senior DBAs run before they reply to a recruiter. A senior production DBA role in Costa Mesa at $160,000 base will sit next to the same role in Seattle at $185,000 and San Francisco at $200,000, and the candidate has already compared all three on a spreadsheet before the first call. Our salary benchmark tool covers metro-level adjustments for DBA roles. Coastal premiums run 15% to 22%. Midwest metros run 10% to 18% below the coastal benchmark. Atlanta, Austin, and Denver run roughly 5% to 10% below.

Certifications That Move the Needle

Certifications matter more for DBA hiring than for SRE or platform engineer roles. The data tier still rewards credentialing because the operational consequences of a bad hire are immediate. Here is what each credential actually tells you when it appears on a resume.

Oracle Certified Professional (OCP) Database Administrator. The most useful Oracle credential. Signals the candidate has worked through the engine end-to-end: instance architecture, backup-and-recovery with RMAN, performance tuning, and the optimizer behavior that trips people up in the exam and in production. Worth requiring at the mid-level for Oracle-heavy roles. The current track is 19c OCP, with 23c emerging. A resume listing 12c OCP is dated but still meaningful for the fundamentals.

Microsoft Certified: Azure Database Administrator Associate. Current Microsoft credential for SQL Server in Azure environments. Replaced the older MCSA path that retired in 2021. Reasonable signal for SQL Server DBAs on the Microsoft cloud. Pair with hands-on Azure SQL Managed Instance experience for senior roles.

AWS Certified Database Specialty. Worth listing as preferred for AWS-heavy DBA roles. Covers RDS, Aurora, DynamoDB, ElastiCache, and the operational AWS database surface. Stronger signal than the generic AWS Solutions Architect for DBA hiring specifically. Less universally useful than OCP for Oracle DBAs.

EnterpriseDB Postgres Professional or Certified Postgres DBA. The credentials that finally exist for Postgres at a serious level. EDB credentials carry weight at Postgres-heavy shops. The Postgres community has resisted formal certification longer than Oracle and Microsoft, so a Postgres DBA without certification but with five years of production experience is still the norm.

Google Professional Cloud Database Engineer. Targets GCP-native database work, including AlloyDB and Cloud SQL. Useful for DBAs at companies that adopted BigQuery and Spanner early. Less universally useful than the AWS Database Specialty because the GCP DBA market is smaller in absolute terms.

One pattern from recent placements: DBA candidates with OCP plus 7+ years of hands-on production work close offers about 18% faster than candidates with either signal alone. The credential filters the pile. The production experience filters the offer. Skip required-cert language entirely at the principal tier. At that level the resume reads cleaner without it, and the work history carries the credentialing weight on its own.

Interview Loop That Actually Screens for DBA Capability

The single most common interview-loop mistake we see is running a DBA candidate through three software-engineering coding rounds and one cursory database conversation. That loop screens for backend engineer. It does not screen for DBA.

One SQL and query-tuning round. Sufficient for most DBA roles, and usually more useful than asking the candidate to invert a binary tree. Give them a real slow query from your production system, anonymized if needed, with the plan attached. Ask them to read the plan, hypothesize where the optimizer went wrong, and walk through the fix. Strong DBAs will ask about the index landscape, the statistics state, the row counts, and the constraints before suggesting anything. Watch the questions. The interview is half the signal.

One backup-and-recovery design round. Pick a real production system. Ask the candidate to design the backup-and-recovery posture for it: RPO and RTO targets, full and incremental schedules, the restore drill cadence, the cross-region replication topology, and the data-loss windows the design actually buys. This is where senior DBA candidates separate themselves from senior backend candidates who happen to know SQL. The wrong answer is a textbook recitation. The right answer starts with questions about the business consequence of data loss.

One incident-response or post-mortem round. Walk the candidate through a real anonymized database incident from your environment, or a public post-mortem from a known company. Ask how they would have run the incident. Ask what action items they would write. Ask what they would have caught earlier. Strong candidates have opinions about blameless culture, the dynamics of getting an action item from “in someone’s queue” to actually shipped, and the way a post-mortem for a data incident differs from one for a service outage.

One platform and IaC round. Especially for cloud platform DBA roles. Walk through how the candidate would build the Terraform module that provisions your database tier, including state isolation, secret management, parameter group design, and the upgrade testing pipeline. This is the round that splits production DBAs from cloud platform DBAs. If the candidate’s whole answer is “I would write a Terraform module,” dig deeper. If their answer starts with “what is the deployment model and who else uses these resources,” promote.

One organizational fit round. Hiring manager plus a senior engineer from the team. The conversation should cover on-call expectations specifically, the team’s appetite for change management rigor, how data work fits into the broader platform roadmap, and what the first 90 days actually look like. Skip the trivia questions. Use the time to make sure both sides understand the working agreement.

Five rounds. One coding round at most, and even that round is optional for senior production DBAs whose day rarely involves writing application code. Reframing the loop from coding-heavy to database-heavy is the single biggest change a hiring manager can make on a DBA search. Candidates walk out more accurately calibrated. Offers tend to land higher because the strongest data-tier engineers feel evaluated for the work they actually do. Across the 17 DBA placements our IT desk closed in 2025 and into 2026, searches where the client used the loop above had an accepted-offer rate of 74%, against 48% on coding-heavy loops. Not enough to generalize beyond our own desk. Enough to copy the shape.

Common Questions Hiring Managers Ask Us About DBA Searches

Is the DBA role going away as cloud-managed databases take over?

The work is moving up the stack, not disappearing. The classic patch-and-tape DBA role is shrinking. The cloud platform DBA and data-engineer-with-DBA-responsibilities roles are growing fast enough to more than replace it.

Managed services like RDS and Cloud SQL absorbed the OS-patching and backup-script work that used to fill a junior DBA’s week. What is left, and growing, is harder: schema and partitioning design at scale, cost optimization, replication topology, data security architecture, and the migration projects that retire the last on-prem instance. Companies that thought they could fire their DBAs after moving to RDS learned within 18 months that they could not, because they had simply traded the operational toil that ran a server room overnight for architectural debt that nobody on the application team was equipped to address during business hours either.

How long should a DBA search realistically take?

40 to 55 days for mid-level production DBA in non-coastal metros when the JD names the engine and posts the comp band. Senior cloud platform DBA runs 55 to 85 days. Data engineer with DBA duties runs longer because the pool is smaller.

The variance is mostly JD quality and engine specificity. Searches where the loop is coding-heavy and the on-call disclosure is buried run 100-plus days. Searches where the loop is database-focused and the comp band is posted close in our 40-day median. Geography matters less than the JD shape.

DBA versus data engineer versus database architect, where is the real line?

DBA owns the running system. Database architect designs schemas and topologies, often without operating them. Data engineer owns pipelines and warehouses. The three overlap, and the line you draw depends on how your org evolved.

At most companies under 100 engineers, one strong DBA does some architecture work and some pipeline work. At companies above that size the roles separate. The fix when the titles get confused is to write the responsibilities clearly. Specify what the role owns end-to-end versus what it advises on. Specify whether the role builds ETL or just operates the database that ETL writes to.

What is reasonable on-call expectation for a DBA in 2026?

One week of primary rotation every three to four weeks is the median we see. More than every other week is a retention risk. Business-hours-only is becoming more common for cloud platform DBA roles where the runtime tier sits behind a managed service.

An explicit on-call premium changes acceptance behavior too. Companies that pay even a modest pager differential close DBA roles faster and retain those hires longer than companies that bundle on-call into base. The dollar amount matters less than the signal: that leadership has put a number on what the rotation costs the person who carries it.

Contract, contract-to-hire, or direct hire for DBA roles?

Direct hire for most production and cloud platform DBA roles. The DBA compounds value through institutional knowledge of your specific data model, and contract DBAs rarely get deep enough fast enough to be net-positive on long-term reliability work.

Exceptions: heterogeneous migration projects with a hard deadline (Oracle to Postgres, on-prem to cloud), post-incident remediation work scoped at three to six months, or capacity work where the DBA is essentially functioning as a senior platform engineer with database focus. Those fit contract or contract-to-hire well. Hourly rates for senior DBA contractors run $115 to $165 in 2026 depending on metro and engine. Our direct hire staffing model applies for the long-term DBA hire, and we will recommend one over the other on intake.

Should I include salary in the DBA posting?

Post the range. Twelve states now require it by law, postings with bands close 24% faster across our IT desk, and a $35K spread is plenty to signal the level without locking you into a number.

The pushback we hear is usually about negotiation leverage. Hiring managers worry that posting a number weakens the eventual offer. The reality on our desk runs the other way: a posting without a range reads to senior DBAs as a posting where someone in HR locked the comp before the spec was scoped, and the resumes that drip in from that posting trend a half-tier below what the JD describes. Real range, real applicant pool, fewer wasted intro calls on both sides.

Where does a DBA actually fit in our org chart?

Production DBAs usually report to infrastructure or operations. Cloud platform DBAs report to platform engineering. Data engineers with DBA duties report through the data or analytics organization.

Reporting line shapes retention. A DBA who reports up through operations but has no influence into the application roadmap ends up either scrapping for budget or quietly job-searching by month nine. The clients that retain senior DBAs past the two-year mark are the ones that route the database conversation into architectural review by default, not by special invitation.

Bringing in Outside Help

If the search has been open more than 60 days, or if the JD has been rewritten twice without the loop changing, the bottleneck is rarely the candidate pool. It is the spec. KORE1’s IT staffing services practice runs DBA searches across database administration staffing at all three archetypes described above. Our 92% twelve-month retention rate on placed DBAs is the metric we watch most carefully on this role, because DBA retention is fragile and the cost of a mis-hire is six months of compound data-tier debt that nobody on the application team can address. We also run Oracle developer staffing and Snowflake engineer staffing when the real need is the engine specialist rather than the operational DBA, and we will tell you which one to hire on intake.

If a second set of eyes on a draft JD would help, or if scoping the right archetype before the requisition opens would save a rewrite cycle, reach out to our team. The intake call runs short, and the typical outcome is a tighter spec, a more honest comp range, and a search that closes weeks earlier than it would have otherwise.