The report arrives. There are posts, campaigns and a graph heading upwards. Yet the question remains: what has actually changed for the business? When your marketing agency is not delivering the results you expected, investigate before deciding whether you need another agency, another plan, or a conversation everyone has been postponing.
We are not going to assume that the agency is doing a poor job or that the client is asking for too much. Without reviewing objectives, agreements and evidence, neither conclusion is justified. Our approach separates the work delivered, its quality and its commercial effect. Those are three different questions, even when they share a presentation slide.
Start with the outcome you actually agreed
Greater visibility, more customers and a stronger brand are not interchangeable objectives. Revisit the proposal and identify the agreed outcome, intended audience, resources and timeframe. Then ask which indicator shows progress towards it. If nobody can explain that clearly, you have found the first conversation to have.
Imagine a hotel commissioning content and social management to improve its presentation, but judging the service entirely by direct bookings. This hypothetical example suggests a possible mismatch. Review whether the scope included acquisition, booking measurement and conversion optimisation. That does not excuse weak content; it distinguishes the commissioned work from what the business now needs.
The opposite situation is possible too: an agency agreed to work towards commercial opportunities but reports only the number of posts produced. Asking how the activity relates to the objective is reasonable. A delivery list is not a substitute for an explanation.
Check whether you are looking at the same evidence
Before debating a percentage, agree what the report counts. Does a conversion mean a purchase, form submission or button click? Are duplicate enquiries included? Have cancellations been removed? Does the reporting period match the sales records? You do not have to solve the entire analytics setup in one meeting, but identify the differences that could change the conclusion.
Google Analytics provides attribution model and lookback-window settings that affect how credit is assigned to interactions. Its attribution documentation explains those options. Our recommendation is therefore to compare definitions and configuration before treating contradictory reports as proof of poor management.
Choose a small sample of real transactions or enquiries and trace what the available records show. The aim is not to reconstruct information that was never collected. It is to establish whether the actions were recorded as expected and where uncertainty remains. Write down the gaps rather than filling them with a convenient explanation.
Separate acquisition problems from conversion problems
We suggest reviewing the journey in stages: exposure, relevant visit, qualified enquiry, proposal and sale. Your business may need different stages, but an explicit sequence helps locate the bottleneck. Saying that marketing is not working is too broad to identify the next change.
If few suitable people arrive, review the message, distribution, searches and intended audience. If interested visitors do not progress, examine the offer, evidence, information and contact experience. Where qualified enquiries fail to become proposals, investigate follow-up and the team’s capacity to handle them.
The agency should explain which parts of that journey it manages, what it needs from the client and what depends on other suppliers. The client should provide the commercial evidence needed to assess its part. Shared responsibility does not mean nobody is accountable. It means putting a name against each responsibility.
Examine the offer you are asking people to buy
Before commissioning more traffic, look closely at the destination page. Is the service understandable? Does the advertisement promise the same thing the visitor finds? Is there availability? Does the offer suit the intended audience and its circumstances? These are business questions, not details to leave outside the marketing discussion.
Consider a hypothetical weekend-package campaign whose landing page neither explains the inclusions nor lets somebody check dates. Changing the targeting might be a test worth considering, but it will not resolve those omissions. First make the offer understandable and usable.
Review objections recorded by sales or reception as well. Do not publish personal information or turn isolated anecdotes into statistics. Use the observations to develop hypotheses that can be investigated. A recurring comment deserves attention, not an automatic conclusion about every potential customer.
Ask for the reasoning behind decisions
Ask the agency what it observed, what it changed, why it made the change and what it expects to learn. That sequence is more useful than a long list of adjustments without context. It should also distinguish implemented work from proposals and items still awaiting approval.
A practical review might take three recent decisions and trace them to the evidence available so far. Where there is not enough data yet, that should be stated. If a test disappointed, discuss what was learned and what follows. Showing only successes makes it difficult to assess the team’s judgement.
Check whether blockers have a named owner and a next action. Material awaiting the client’s approval should not be presented as completed work. Equally, an agency task that was not delivered should not disappear behind a general request for patience.
Distinguish reasonable timescales from waiting without a plan
Different changes need different evaluation periods. In its guidance on content improvements and core updates, Google explains that effects can take time and are not guaranteed. That does not establish that every delay is normal, or validate a particular commercial promise.
Our proposal is to agree observable milestones: assessment completed, changes published, measurement checked and a review date set. The timing of a commercial outcome may be uncertain. The existence of the work and an explanation of its dependencies should not be confused with that uncertainty.
Consider your own purchase cycle too. A service requiring several meetings cannot reasonably be judged only by sales closed immediately after a campaign launches. You can still review whether relevant opportunities are appearing and progressing through the agreed process.
Run a meeting that ends in decisions
Bring the original objective, comparable data, a sample of enquiries and the delivery record. Ask for a prioritised diagnosis rather than a defence of every activity. The outcome should be a short plan containing the main observed problem, the improvement hypothesis, a specific action, an owner and a review date.
Keep the priorities limited. Changing the audience, message, page and follow-up simultaneously may make the result harder to interpret. Isolating every variable will not always be practical, but the changes and that limitation should be documented.
Agree which evidence would support continuing, adjusting or reconsidering the relationship. You do not have to demand a sales guarantee in order to expect clarity, verifiable work and a professional explanation. Those expectations can be firm without turning the meeting into an argument.
When a change of agency deserves consideration
We would consider it reasonable when there is a persistent mismatch of expertise, a lack of transparency or documented failure to deliver agreed work. Base the decision on evidence and the business’s current needs, not merely a disappointing month or another supplier’s most impressive promise.
Before a transition, organise authorised access, assets, campaigns, content and documentation. Decide what will continue and what needs reassessment. Replacing the contact person without correcting the objective or measurement can move the same problem into a different meeting.
At Isaura, we prefer to understand the complete journey, from strategy and execution to the website and analytics. For a second perspective, tell us what you expected to achieve and what is happening. The first job is to define the problem, not to promise that a different logo on the report will solve it.
