Most agency shortlists in the UAE are built from capability decks. That is understandable and usually unhelpful. The same logos, platforms, dashboards, and claims appear in almost every deck, while the expensive part of the engagement remains invisible: who makes the decision when media, creative, the website, and the CRM point in different directions?
That is the decision this article is designed to help with. I am not offering a ranking of UAE agencies. I am giving you the diligence method I would use if I were appointing a partner to own a growth system rather than a channel.
My view is simple: do not choose a MarTech partner by counting services. Choose by inspecting ownership at the handoffs.
A strong MarTech partner is not the one with the longest tool list. It is the one that can explain which decision each system exists to improve.A service list tells you what an agency can invoice. A decision trail tells you whether it can operate your growth.
What commercial problem are you actually trying to solve?
“MarTech” is useful shorthand, but it does not define a standard operating model. One firm may be a media agency with analytics support. Another may be a software team that implements a CRM. A third may own acquisition, landing pages, tracking, creative, and automation together.
Before comparing them, write down the problem you actually need solved:
- Paid media is producing leads, but sales rejects too many of them.
- Campaign learning is slow because page changes sit in another vendor's backlog.
- Revenue reporting changes depending on which platform is open.
- Creative production is disconnected from live campaign evidence.
- Follow-up is inconsistent after the lead enters the CRM.
Now ask which parts of that problem the proposed partner can directly change. A broad agency is not automatically better. If the bottleneck is a well-defined Salesforce implementation, a specialist may be the sharper appointment. If the bottleneck crosses media, the landing experience, event tracking, and follow-up, appointing four specialists without an internal integrator recreates the problem.
Which four handoffs decide whether the plan survives?
At True North, performance marketing, web and software development, creative, and AI automation work from one commercial target. That structure exists because these four handoffs repeatedly determine whether a plan survives execution.
The partner does not need to perform every task with permanent in-house staff. It does need to name who remains accountable when a specialist or production partner is involved. “We have a network” is not an answer to ownership.
What decision evidence should you actually ask for?
Case studies show that a result happened. They rarely show whether the operating method will transfer to your business. During diligence, I would ask for three redacted artifacts.
1. A decision note
This can be a weekly report, an experiment log, or a short change record. I want to see:
- what the team believed;
- what changed in the account or experience;
- which evidence was reviewed;
- what it decided to stop, hold, or do next;
- what remained uncertain.
The format does not matter. The presence of a decision does.
2. A measurement map
The map should connect a user action to the systems that record it. For lead generation, for example, the team might distinguish a submitted lead, a qualified lead, and a closed lead. Google Analytics publishes recommended events such as generate_lead and qualify_lead, but using a recommended name does not make the underlying business definition correct. The definition still has to be agreed and implemented. Google's event reference is useful for technical consistency; your CRM and sales process remain the commercial source of meaning.
3. An underperforming piece of work
Ask what the team expected, what happened, and what it changed. A partner that can only discuss wins is either withholding the useful part of its experience or has not built a learning process.
Why does attribution need governance, not just software?
Every ad platform, analytics product, and CRM observes a different part of the journey. Disagreement is normal. What matters is whether the team has agreed rules for working through it.
I look for four things:
- Definitions. What counts as a lead, qualified lead, sale, refund, or repeat purchase?
- Change control. Who can edit conversion events, campaign naming, or CRM stages, and where is that recorded?
- Reconciliation. Which report is used for budget decisions, and how are platform-reported conversions compared with finance or CRM outcomes?
- Limitations. Which parts of the journey are not observable, and how does that uncertainty affect the decision?
An agency that promises “perfect attribution” is avoiding the real conversation. A useful partner can tell you which model it is using, why it is adequate for the decision, and where it may be wrong.
How do you test whether an agency's AI claims are real?
“AI-powered” has become as empty as “data-driven.” I do not evaluate it as a feature. I ask four operational questions:
- Which task is assisted or automated?
- Which information may the system access?
- Which person reviews the output before it affects spend, customers, or published work?
- What failure is the workflow designed to catch?
AI can help with research, classification, creative variations, reporting preparation, and internal workflow automation. It can also produce confident errors and accelerate a bad definition. The question is not whether the agency uses AI; it is whether the workflow makes the use inspectable and reversible. We apply the same standard when scoping AI automation inside a marketing system.
What should the proposal say about exclusions and dependencies?
A serious proposal should make the uncomfortable parts visible. If an agency is accountable for paid acquisition but cannot change the website, access the CRM, or receive sales feedback, that limitation belongs in the scope. If the client must supply product photography, legal approval, inventory data, or engineering access, say it before the start date.
I would expect the proposal to name:
This is also where an integrated model should prove its value. Our model is founder-led, with one named lead accountable for the engagement and a shared brief across the four pillars. That does not remove dependencies or uncertainty. It reduces the number of external handoffs needed to act on them. The broader structure is explained on our services page.
When should you choose specialists instead of an integrated partner?
There are situations where I would not recommend one broad partner:
- You have a strong internal growth owner with authority over all vendors.
- The work is a deep, bounded implementation that does not depend on adjacent disciplines.
- Procurement requires separate specialists and your governance is mature enough to connect them.
- The agency's integrated offer is broad on paper but shallow in the area that creates most of your risk.
Integration is not a virtue by itself. It is useful when it shortens the path from evidence to action without lowering specialist quality.
Make the shortlist earn its way forward
Give each shortlisted partner the same commercial problem and ask for the same three artifacts: a decision note, a measurement map, and a reversed or failed test. Then compare the quality of the questions they ask, the boundaries they state, and the owner they put forward.
If you are considering True North, use this standard on us. Review our ecommerce case studies, then ask us to walk you through how each number was measured, inspect how the four disciplines connect, and ask us what we would not control in your scope. If the operating model fits, start the conversation. If a specialist is the better answer, the diligence should make that clear too.
