Measurement work for Downtown Ottawa and Centretown, where the buyer is a procurement process
Downtown and Centretown hold the professional services firms, national associations, consultancies and software vendors that orbit the federal government, and the way those organisations buy and get bought from breaks most of the assumptions a marketing measurement stack ships with. There is frequently no purchase event to instrument, because the decision is a tender, a standing offer or a committee vote months downstream. There is frequently a lawyer or a privacy officer who has to approve a tag before it goes near a page, and who will say no to a third-party pixel on any form that collects personal information. And the audience for the reporting is a board or a committee that meets four times a year.
Why Downtown and Centretown is its own case
Three constraints define the work here, and all three arrive before any campaign question does.
The first is that procurement is not a funnel. When an organisation sells into government or into a large institution, the transaction runs through a solicitation, a vendor of record arrangement or a standing offer, and none of those events happens on a website that anyone can instrument. Optimising toward a contact form in that setting means optimising toward the least serious contact you receive. What can be measured honestly is narrower and more useful: reach against the specific accounts and roles that matter, engagement with the material that actually gets forwarded internally, and a named opportunity opened in the CRM by a salesperson who logs it. Everything downstream of that is a procurement timeline, and pretending a platform can attribute it is not measurement, it is decoration.
The second is that instrumentation has to survive a review. A tag manager is a code deployment path into a production website, and in an organisation with a privacy function it is treated as one. Custom HTML tags get read. Third-party scripts get refused on pages that handle personal information. Consent has to be recorded in a way that can be produced later, not merely collected. Ontario organisations over a certain size also carry accessibility obligations for their public websites, and a landing page built quickly for a campaign is exactly where that gets forgotten. None of this is a reason to measure less. It is a reason to write the tracking plan before the tags, name the data each one collects, and get the review done once rather than three times.
The third is the reporting cadence. A committee that meets quarterly cannot be served by a live dashboard nobody opens between meetings. It needs a document that is generated rather than assembled, with definitions attached, so that the meeting argues about the decision instead of about whose spreadsheet is right.
The work, as it applies here
GA4 & GTM implementation
A documented tracking plan first, tags second: what each tag collects, where it sends it, what legal basis it relies on, and which pages it is forbidden on. Written to be handed to a privacy reviewer rather than reverse engineered by one after the fact.
Conversion attribution audit
Working out what is genuinely measurable when the purchase is a procurement file, and saying plainly which of the numbers currently on the report are inferences rather than observations.
Automated client reporting
A quarterly document that produces itself for a board or a committee, with the metric definitions travelling alongside the numbers, so the same figure means the same thing in March as it did in December.
FAQ: Downtown and Centretown
Our buyers procure through solicitations and standing offers. What is there to optimise toward?
Not a form, because the form is not the decision. What holds up is a short list of measurable proxies chosen deliberately: reach and frequency against the organisations and job functions you actually sell to, engagement with the specific documents that get circulated internally, and opportunities opened in your CRM by a person who logs them. Optimise toward the opportunity, report the proxies as leading indicators, and state on the report that the procurement timeline is not attributable. That last sentence saves more credibility than any model will.
Legal will not allow a third-party tag on a page that collects personal information. What can we still measure?
Most of what you need, if the measurement is designed around the restriction instead of arguing with it. The page can fire a first-party event that records the submission without transmitting the field contents, and the qualified outcome can be sent later from your own systems, server side, once the record has been reviewed. That separates the fact that a conversion happened from the personal information inside it. It also tends to produce a cleaner dataset than the version legal refused, because it stops sending free-text fields to a third party by accident.
Do Ontario accessibility obligations apply to a landing page we built for a campaign?
A public-facing web page is a public-facing web page, and a campaign landing page does not sit outside the obligation because it was built quickly or lives on a marketing subdomain. The practical consequence is that colour contrast, keyboard operability, form labels and a sensible heading order belong in the build checklist rather than in a remediation project after somebody complains. It is also cheaper there. Retrofitting a page template is a rewrite, while getting it right the first time is an afternoon.
Someone submitted an access request asking what we hold about a website visitor. What does that touch?
More systems than the marketing team usually expects, which is why the answer is a preparation problem rather than a response problem. Analytics identifiers, advertising cookies, form submissions, the CRM record, the email platform and any warehouse copy are all potentially in scope, and a request arrives with a clock attached. The work worth doing in advance is an inventory: what is collected, where it lands, how long each copy is kept, and who can delete it. An organisation that has that written down answers in days rather than weeks.
Can we track a conversion that completes on a government portal we do not control?
Not through to completion, and it is better to say so than to model it. You can measure everything up to the handoff, including the click that leaves your site for the portal, and you can instrument the return path if there is one. Beyond that boundary you are relying on the other organisation to tell you, which usually means a manual reconciliation on a schedule rather than a data feed. Design the report around the handoff as the last observed event and label it as such.
How do we report to a board or a committee that only meets quarterly?
With a generated document rather than a live dashboard, and with the definitions printed next to the numbers. A quarterly audience has no memory of the operational context, so a metric that changed definition since the last meeting reads as a result rather than as a change of method. The reporting system should produce the same document every quarter without anyone rebuilding it, which is what makes the trend line credible: the numbers moved, the method did not.
Reading that applies
Also covered by this page
These 11 places raise the same questions as Downtown and Centretown and are answered here rather than on a page of their own. Every one of them is inside the City of Ottawa.
- Tunney's Pasture
- Golden Triangle
- Centretown West
- LeBreton Flats
- Lees Avenue
- New Edinburgh
- Lindenlea
- Manor Park
- Wateridge Village
- Sandy Hill
- Rockcliffe Park
Kodelytics Inc. works remotely from one address in Kanata and has no office, phone line or staff in Downtown and Centretown. This page describes how the work applies to businesses there, and claims nothing else.