Analytics and paid media operations for Kanata North, where the sale closes months after the click
Kanata North is the largest technology park in the country, and the measurement problems that come out of it look nothing like the ones a local service business has. The buyer is a committee. The deal is signed one, two or three quarters after somebody first typed a search. The thing worth counting is not a form submission, it is an opportunity that reached a named stage in a CRM, and the CRM is usually the one system the marketing team does not own. Meanwhile the ad platform is being asked to optimise toward a signal it never receives, and the quarterly deck reports a cost per lead that nobody in the room believes, because the leads it counts and the pipeline finance counts are two different populations.
Why Kanata North is its own case
Three things go wrong here often enough to be predictable, and none of them is a bidding problem.
The first is that the sales cycle outruns the attribution window. Ad platforms attribute a conversion back to a click for a bounded number of days, and that ceiling is measured in weeks while an enterprise software cycle is measured in quarters. A conversion recorded after the window closes is not attributed late, it is not attributed at all, and the campaigns that sourced the largest deals look like the worst performers in the account. The fix is mechanical rather than clever: capture the click identifier on the form, carry it into the CRM as a field on the lead, and upload the qualified stage back as an offline conversion so the platform learns from outcomes instead of from form fills. It only works if somebody owns the field, which is the part that gets skipped.
The second is that the conversion object is in the wrong place. A product-led company puts its real signal in the application database, not on the marketing site: the trial that activated, the workspace that invited a second user, the account that survived thirty days. A sales-led one puts it in the opportunity record. Either way the website event is a proxy, and treating the proxy as the target is how an account ends up optimised toward demo requests from students and competitors. The lead-to-opportunity lifecycle has to be defined before any of it is instrumented, because the definition decides what gets built.
The third is the reporting audience. A vice president of marketing at a Kanata North firm is not asking what the cost per lead was, they are being asked by a board what paid media contributed to pipeline and what it will contribute next quarter. That question cannot be answered inside an ad platform, because platform data expires, is retained on the platform's schedule, and cannot be joined to a deal record. It is answered in a warehouse where the ad spend, the web session, the lead and the opportunity are the same row.
The work, as it applies here
Conversion attribution audit
Reconciling the four systems that disagree in a business-to-business stack: the ad platform, GA4, the CRM and the finance report. The output is a written account of where each number comes from, which of them is allowed to be the answer, and what is genuinely unknowable given the current instrumentation.
GA4 & GTM implementation
Event architecture built around the pipeline rather than around the page: click identifier capture and persistence, enhanced conversions, consent signals, and cross-domain measurement into a documentation site, a trial application or a booking tool on a different hostname.
Marketing data warehouse
A BigQuery layer that outlives the platform retention windows and holds spend, session, lead and opportunity in one place, with metric definitions written down so pipeline sourced and pipeline influenced mean one thing across marketing and finance.
FAQ: Kanata North
Our sales cycle runs six to nine months. Can Google Ads measure that at all?
Not on its own, and not from the click. A conversion has to be recorded inside the platform’s attribution window to be credited to the click that caused it, and that window is far shorter than your cycle. The route that works is to store the click identifier when the lead is created, carry it as a field on the CRM record, and upload the qualified or closed stage back as an offline conversion when it happens. The platform then learns from real outcomes, and the deal is credited to the campaign that sourced it even though the click was three quarters ago.
Should the CRM or GA4 be the source of truth for our pipeline numbers?
The CRM, without much argument, because it is the only one of the two that knows whether a lead became money. GA4 is the right source for behaviour before the form: which channel, which content, which sequence of sessions. The mistake is asking either of them to be the answer to both questions. Write down which system owns which metric, publish that document, and treat any number quoted from the wrong system as an error rather than a discrepancy to be explained again next month.
We are product-led. The signup happens in the app, not on the marketing site. Where does the conversion live?
In the product database, and that is a decision with consequences rather than a detail. If the meaningful event is activation rather than registration, the marketing site cannot observe it, so it has to be sent from the application back to the ad platforms and the analytics property with an identifier that ties it to the original session. That means cross-domain measurement between the marketing hostname and the app hostname, a stable identifier that survives the account creation step, and agreement on what activation actually means before anyone writes the event.
Do we need a data warehouse, or is a dashboard on top of the ad platforms enough?
A dashboard is enough while every number you need lives in one platform and you never need it older than that platform keeps it. A business-to-business firm here usually fails both tests at once: the deal record is in a CRM, the spend is in two or three ad accounts, and the board wants a year-over-year view that outlives the platforms’ own retention. The honest test is whether anyone is currently pasting exports into a spreadsheet to answer a recurring question. If they are, the warehouse is already being built by hand, badly.
Our parent company is abroad and runs the ad accounts. What can a Kanata team actually change?
More than most teams assume, and the constraint is usually access rather than authority. Even where campaign management sits elsewhere, the tracking plan, the conversion definitions, the CRM fields and the reporting layer are local decisions, and they are the ones that determine whether the head office numbers mean anything for this market. The first thing worth establishing is read access to the account and a named conversion action that reflects the Canadian pipeline, so the regional result stops being an inference drawn from a global blended figure.
How do we report paid media to a board that thinks in pipeline, not clicks?
Report sourced pipeline, influenced pipeline and cost per opportunity, define all three in writing, and never change a definition mid-year without saying so on the slide. Clicks, impressions and cost per lead belong in an appendix that supports the operating team, not in the board deck. The reason this is a data engineering job rather than a slide design job is that sourced and influenced pipeline both require the ad spend and the opportunity record to sit in the same table, which is exactly what the platforms cannot do for you.
Reading that applies
Also covered by this page
These 10 places raise the same questions as Kanata North and are answered here rather than on a page of their own. Every one of them is inside the City of Ottawa.
- Beaverbrook
- Katimavik
- Glen Cairn
- Morgan's Grant
- Marchwood Lakeside
- Briarbrook
- Kanata Town Centre
- South March
- Kanata Highlands
- Kanata Lakes
Kodelytics Inc. works remotely from one address in Kanata and has no office, phone line or staff in Kanata North. This page describes how the work applies to businesses there, and claims nothing else.