A large share of the organisations buying SEO in this city cannot simply pick a supplier they like. Health systems, universities, hospitals, law firms and public bodies run a procurement process, and that process has its own gravity. It rewards vendors who are good at responding to procurement, which is a different skill from the one you are trying to buy. If you are the person writing the request, the leverage is almost entirely in how you write it.
Institutional buyers in Philadelphia choose SEO suppliers through procurement
The standard template asks for company history, references, insurance, a methodology section and a price. Every capable firm and every incapable firm can answer all of that. The result is a stack of documents that look interchangeable, a scoring sheet that cannot separate them on substance, and a decision made on price or on who wrote the most confident methodology.
The fix is to ask for work rather than for claims. Give every respondent the same small, real problem from your own site and ask for a written analysis of it, with a page limit. It costs each of them a few hours, it is impossible to answer with boilerplate, and it produces the only genuinely comparable material in the whole exercise. Firms that decline to do it have told you what their proposals are made of.
A badly written request selects for proposal writing, not SEO skill
Three habits do most of the damage. Asking for a list of services, which every firm has. Asking for case studies, which selects for whoever has the largest clients rather than the most relevant experience. And scoring methodology sections, which rewards fluent generic prose.
Replace them with specifics. Name the pages or sections you believe are underperforming and ask what the respondent thinks is wrong with them. Ask who would be assigned, by name and role, and require that those people appear at any presentation. Ask what they would recommend not doing. And ask for one reference from a client who left, which is the single most informative question available and almost never appears in a template.
Subdomain sprawl is the most expensive problem on a large site
Institutions here run enormous web estates that grew by accretion: a separate site for each department, school, clinic or practice group, several content systems, a few forgotten microsites from campaigns long finished, and the same information published in four places with slightly different wording. Every one of those properties competes with the others.
The valuable SEO work in this environment is almost never writing more. It is consolidation, redirect strategy, deciding which entity owns which topic, and building the governance to keep it that way. That work is unglamorous, politically difficult and produces the largest returns available to a big organisation. Ask respondents directly how they would approach a consolidation, and whether they have done one at your scale. The answers separate the field quickly.
Scoring criteria that reward the right things
Publish the weighting in the request. If the analysis exercise is worth the most, say so, and make price a smaller share than the instinct of a procurement office would suggest. Cheap SEO on a large institutional site is not a saving, because the cost of a botched migration dwarfs any fee difference.
Insist that the price is broken into analysis, production and ongoing programme management, rather than presented as a single annual figure. It makes the responses comparable, it makes later renegotiation possible, and it exposes the firm whose attractive total quietly excludes all the writing. If running a comparable process is the part you would rather not build from scratch, set out the requirement once and ask for replies in a fixed shape.
Decide who can publish and who can change an address
In a decentralised organisation the biggest risk to a search programme is not the agency. It is the forty people with publishing rights who can create a duplicate page, retire a URL without a redirect, or launch a microsite on a new domain because it was easier than negotiating with the central team.
So make governance part of the scope rather than an afterthought. Who approves a new section. What the rule is for retiring a page. Who holds the redirect map and reviews it before each release. An SEO engagement without that structure will spend its budget repairing damage created faster than it can be fixed.
Reporting to a committee without losing the point
Institutional reporting drifts toward dashboards because dashboards are safe. Resist it. Agree one primary measure that the organisation genuinely cares about, expressed in its own language: appointment requests, applications started, enquiries from a named service line, referrals. Report it monthly with its cost, and keep visibility and position data behind it as diagnostics.
Then hold a standing quarterly question: what did we stop. Large programmes accrete activity because every initiative has a sponsor and nothing is politically easy to cancel. A review that only adds is not a review, and after two years a meaningful share of the budget is usually maintaining things nobody has evaluated since they launched.
Accounts and property matter in an SEO contract too. Analytics, the search console, the tag container, the content systems and the domain registrations must sit with the institution, with vendors granted access, and every document produced under the contract should transfer with editable source files. For a wider view of how these suppliers are organised and priced, start from the overview of SEO agencies. If the constraint is the platform rather than the pages, talk to local software teams or development firms working at this scale, and if you are widening the field, compare against agencies in the larger market to the north or firms serving comparable institutions.