Ads, conversion work and analytics act as one system only if the numbers under them can be trusted and read. This post sets out the checks we run before paid spend goes up, the account structure that keeps results readable, and where testing fits. The example is our Google Ads work for Bruc Bond in Singapore and the UK.
Why the order matters more than the tools
Most advice on joining up paid media, CRO and analytics is about shared dashboards and weekly meetings. Those help, but the usual failure is sequence. Budget is raised on a conversion count nobody checked, a landing page is tested on the wrong traffic, and the report blends both into one number.
We run the three in a fixed order. First the measurement is made true, then the account is built so each result can be read on its own, and only then are pages tested, where traffic supports an honest test. Skipping a step does not save time; it moves the cost into decisions made on the wrong figures.
- 01Measurement made trueConversions checked against real orders and CRM records
- 02Account built to be readSeparate campaigns per market and product, business intent only
- 03Pages testedOnly where traffic supports an honest test
The measurement checks we run before spend scales
Our paid media retainer says tracking is set up and verified before spend scales, and that line is in the scope for a reason. These are the checks behind it. Each one has a pass condition you can test yourself.
| Check | What we look at | It passes when |
|---|---|---|
| Conversion definition | What the account counts as a conversion | The primary conversion is an order, a qualified lead or a signed account, not a page view or a form opened |
| Real-world match | Platform conversions against actual orders or CRM records for the same days | Every gap between the two is listed and explained |
| Consent and tagging | How tags fire when a visitor declines cookies, and whether server-side tagging is in place | The setup is written down and tested on the live site, not assumed from the tag manager |
| Offline outcomes | Whether sales that close later, by phone or in a CRM, reach the ad platform | They are imported, or the report says plainly that they are missing |
| Two views side by side | Platform-reported results beside analytics results | Both appear in the same report, so nobody has to choose silently between them |
None of this is exotic. It is skipped because each check sits between two owners: the ads specialist trusts the tag, the analyst trusts the platform, and the developer was never asked. When one team holds all three, the check has an owner.
Bruc Bond: building the account so results can be read
Bruc Bond provides corporate accounts, cross-border payments and business banking, with a software platform built on the same capabilities. We ran its Google Ads campaigns, aimed mainly at Singapore and the United Kingdom. The website is the client’s own; our work was the paid search account.
Three things made the account hard to read before a single bid changed. Two markets with different search behaviour. A financial offer where the searcher must be a business, not a consumer. And two products, banking services and the platform, competing for one budget.
We answered each with structure rather than with reporting:
- A campaign per market. Singapore and the United Kingdom run separately, so budgets, bids and messages follow each market instead of averaging across both.
- Business intent only. Keywords and negative keywords were built around what companies search when they need corporate banking or international payments, with consumer queries kept out.
- Each product on its own. Banking services and the platform run as distinct campaigns with their own messages, so each product’s performance can be judged without the other blurring it.
The analytics lesson is that a report can only separate what the account separates. If two markets share a campaign, no dashboard will tell you which one is paying. The full structure is in the Bruc Bond case study.
Where conversion work comes in
Conversion testing is the third step because it depends on the first two. A test on a landing page is only as good as the traffic reaching it and the conversion it counts. With measurement checked and campaigns split, the conversion team knows which audience a page is serving and what a win means.
We choose between testing and direct fixes on traffic, not preference. An A/B test runs only where the page gets enough conversions to reach significance in a sensible time. Below that, our conversion team works from recordings, funnel analysis and on-site surveys, ships fixes directly and logs every result as win, lose or inconclusive.
The handoff runs both ways. What the ads team learns about which message wins goes into the landing page brief, and what the page data shows about where visitors drop out goes into the next round of ad copy. That loop is the practical meaning of having all three on one roadmap.
When this is not the right answer
The full sequence is not worth it for every account. It is overkill when:
- Spend is small and the account is a single campaign. A tested conversion tag and a monthly look at search terms may be all it needs.
- The business closes a handful of large deals a quarter. Conversion counts will never support an A/B test, so the money is better spent on the offline import and on the sales handoff.
- Tracking is broken and nobody can fix the site soon. Raising budget first is a mistake, but so is a long testing programme. Fix the measurement, then decide.
In those cases we say so, and sometimes the answer is a one-off audit rather than a retainer. We wrote about the same judgement in why our PPC team sometimes recommends you spend less.
Written by
Shweta Agarwal
Chief Operating Officer, Vinove
Shweta runs the operating side of the group, holding delivery, process and accountability consistent across the group. Her focus is on the systems that let teams grow without losing the standard they started with.
LinkedIn →
