The dashboard looks great. Impressions are climbing, install numbers are up, and cost-per-install is sitting right where the projection said it should be. Then someone runs the retention numbers, and the picture changes entirely. Day-7 is soft. Day-30 is worse. The users who looked cheap to acquire aren’t doing anything with the product, and the return on ad spend, measured against actual revenue, is a fraction of what the surface metrics suggested.

This is the gap between vanity metrics and real ROAS, and it’s where a significant portion of user acquisition budget quietly disappears. Nielsen’s Annual Marketing Report found that while 84% of marketers are highly confident in their ROI measurement, only 38% actually evaluate holistic ROI by measuring traditional and digital channels together. The confidence is real. The completeness of the picture, less so. 

Dragalinos Limited delivers full-service user acquisition for digital platforms – managing the strategy, the channel mix, and the performance measurement, while bringing in channel-dedicated external experts to execute each traffic source to the depth it requires. The model is built around a specific conviction: that the measurement problem and the staffing problem in UA are the same problem viewed from different angles. What follows is how Dragalinos approaches both.

What Vanity Metrics Actually Measure

Vanity metrics are the numbers that look like progress without actually being progress. In user acquisition, that usually means install volume, cost-per-install, and click-through rate – three figures that move in visible, reportable ways and carry a surface logic of their own. Lower CPI feels like efficiency. More installs feel like growth. Higher CTR feels like resonance. None of that feeling is wrong exactly — the problem is that none of it answers the question that matters, which is whether the people coming in are staying and doing something with the product.

None of these is useless. They’re leading indicators. The problem comes when they’re used as success metrics rather than as proxies for something that needs to be measured further downstream. A low CPI means nothing if the users acquired at that cost leave before completing onboarding. A high click-through rate means nothing if the traffic it’s generating is composed of bot clicks or casual browsers with no intent to engage.

According to Dragalinos, the shift from vanity metric optimization to real ROAS optimization is one of the most consistent improvements available to platforms that feel like their UA spend isn’t working. The campaigns are often fine. The measurement framework is what’s broken — it’s rewarding the wrong outcomes and creating pressure to optimize for metrics that don’t reflect business value. Dragalinos has seen this same pattern across enough platforms to treat it as a structural issue rather than an individual campaign problem.

How Vanity Metrics Get Entrenched

The persistence of vanity metrics in UA reporting isn’t usually a mistake – it’s a structural outcome of how campaigns get evaluated. Short reporting cycles favor metrics that move quickly and look good on a weekly basis. Install volume, CPI, and click volume all update in near-real time. Retention curves, lifetime value, and true ROAS take weeks or months to develop.

When performance is being assessed week-over-week, the incentive naturally shifts toward metrics that give weekly feedback. Dragalinos Limited has observed that teams operating under short reporting cycles tend to optimize toward whatever metric moves in the right direction by Friday, which is almost never LTV or real ROAS, because those figures aren’t yet available.

The fix requires extending the reporting time horizon and anchoring campaign evaluation to the metrics that reflect actual business outcomes, even when that means accepting less frequent feedback. It also means being willing to break down performance at the channel and creative level rather than stopping at the campaign aggregate – because a poor-quality traffic source can look fine in aggregate data while quietly dragging down everything downstream.

Traffic Quality as the Foundation of Real ROAS

Traffic quality is the property that determines whether the users a campaign brings in are worth what was paid to acquire them. It sounds like a secondary concern — something to think about after the installs are coming in at target cost. In practice, it’s the primary variable in whether a UA campaign produces real ROAS or just activity. Dragalinos Limited treats traffic quality as the first question in campaign evaluation, not an afterthought once the numbers are already in.

Low-quality traffic comes from many sources. Click fraud and install fraud are the most obvious, but they’re not the only ones. Incentivized installs from users who have no genuine interest in the product inflate volume with users who will never engage. Broad targeting that pulls in users who fit a demographic profile but don’t have the behavioral intent to engage with the specific product. Creative that attracts clicks from curiosity rather than genuine purchase consideration.

What all of these have in common is that they produce installs that look normal in a dashboard until the downstream behavior is examined. The install metric doesn’t know the difference between a user who downloaded the app because they’re going to use it and a user who downloaded it because an incentive program paid them to.

Dragalinos Limited’s approach to UA is built around traffic quality as a non-negotiable foundation. Every channel in the acquisition stack gets evaluated not just on the volume and cost of the installs it delivers, but on the quality of behavior those installs produce downstream – retention curves, engagement depth, conversion to revenue-generating activity. Channels that can’t sustain quality benchmarks don’t get budget, regardless of how attractive their surface metrics look. Dragalinos treats this as a first principle, not a refinement.

The problem with enforcing that standard is that it requires genuine channel-level expertise to act on. Spotting the specific creative pattern on TikTok that’s pulling in curious clickers rather than intent-driven users is a different skill from spotting the same problem on Meta. Identifying why Apple Search Ads traffic is converting at install but dropping off at activation requires a different diagnostic lens than the same problem on Google App Campaigns. A team that’s spread across all of these simultaneously will almost always default to aggregate metrics — not because they’re careless, but because the channel-level depth required to go further simply isn’t available when one person is doing five channels at once. That’s what connects the measurement problem to the staffing question.

Why Channel Specialists Outperform Generalists

The channel structure is where that conviction becomes concrete. A generalist UA manager who runs campaigns across TikTok, Meta, Google App Campaigns, Apple Search Ads, and YouTube simultaneously manages five fundamentally different advertising environments – distinct creative requirements, audience mechanics, optimization logic, fraud risk profiles, all running at the same time inside the same person’s attention. Dragalinos has found that this is where UA quality problems typically originate, long before they show up in any dashboard.

The generalist approach has an obvious appeal. It’s simpler to manage, requires fewer external relationships, and feels more operationally unified. The difficulty is that each of those channels has a depth of specialization that one person running all of them simultaneously can’t fully reach. TikTok creative that drives genuine intent behaves differently from Meta creative doing the same job – not slightly differently, but in ways that require different intuitions, different testing approaches, and different pattern recognition developed over time. The optimization signals on Google App Campaigns work differently from those on Apple Search Ads. A manager covering all of it is doing each of them at a fraction of the depth a dedicated specialist can reach, and the output reflects that. As Dragalinos Limited has observed across client campaigns, the quality gap between a generalist and a channel-dedicated specialist tends to be invisible at the CPI level and very visible at the retention level.

Dragalinos structures UA around channel-dedicated specialists – one person or team per traffic source, focused entirely on that channel and nothing else. The person running TikTok UA isn’t also running Meta. They’re not context-switching between platforms with fundamentally different creative logic and optimization mechanics. What they know about TikTok comes from running TikTok campaigns continuously, across enough volume and variation to develop the kind of intuition that generalists working across five channels simultaneously simply can’t accumulate at the same depth. Meta, YouTube, and Google App Campaigns, Apple Search Ads – each gets the same treatment. Dragalinos has found that this consistently produces better quality metrics and lower cost-per-retained-user than generalist configurations, not by a small margin.

Why External Specialists Outperform In-House Generalists

The staffing model matters here. Hiring channel specialists in-house creates a team that’s expensive to maintain regardless of campaign activity. A TikTok specialist on salary during a quarter where TikTok is deprioritized is overhead without proportional output – and that situation comes up more often than it might seem, because channel mix shifts constantly with platform performance, seasonality, and budget cycles. Multiply that across five or six channels, and the in-house model starts carrying high fixed costs just to maintain coverage.

The alternative Dragalinos Limited has developed is working with external specialists for each channel — whether that’s a freelance professional or a specialist contractor company focused on a single traffic source. The distinction matters less than the principle: whoever handles TikTok UA does only TikTok UA. A specialist of that kind isn’t accumulating expertise in isolation. They’re running TikTok campaigns continuously across multiple clients, which means they’re seeing more data, more edge cases, and more creative variation than any single in-house role would ever generate. The same applies to a specialist contractor company that focuses exclusively on, say, Apple Search Ads — their entire organizational knowledge is concentrated on one channel, not spread across a portfolio of services.

This model has operational advantages that go beyond just expertise depth. The cost structure flexes with campaign activity – when a channel is less active, the engagement scales down rather than sitting as fixed overhead. The performance bar stays high because external specialists operate in a competitive market — those whose results don’t justify their rate simply don’t get repeat work. And the channel knowledge stays genuinely current, because specialists whose entire commercial value is tied to one platform have strong incentives to stay at the frontier of what’s working there, not just maintain what worked six months ago. Dragalinos Limited has found that this competitive dynamic produces a quality floor that in-house teams – where underperformance is more insulated from consequences – rarely sustain at the same level.

Dragalinos Limited uses this model specifically to avoid the quality dilution that comes from asking one person or one team to cover too many things at once. The result is a UA operation where each channel gets the depth of specialization it actually requires – without the overhead of maintaining a full in-house team for every traffic source, regardless of what’s running that month.

Measuring Real ROAS Across a Specialist-Driven Stack

The practical challenge of running a multi-specialist UA stack is maintaining coherent measurement across channels that are being optimized by different people with different tooling and reporting. This is where a lot of the operational value of the Dragalinos Limited model comes from – the integration layer that sits above the individual channel specialists and maintains a consistent view of real ROAS across the full acquisition stack.

Each specialist is accountable for channel-level metrics: quality-adjusted install cost, retention cohorts by creative, and the conversion rates at each step of the funnel for traffic originating from their channel. Those metrics feed into an integrated performance view that allows Dragalinos to see true ROAS – not just cost-per-install by channel, but revenue generated per dollar spent on acquisition, segmented by channel and tracked through the full user lifecycle. This integrated view is what turns individual channel optimization into system-level performance improvement.

This measurement architecture is what makes the specialist model produce better outcomes than a generalist approach, not just better individual channel performance. The quality data from each specialist feeds a system-level optimization process that allocates budget toward the channels and creative approaches that are producing the best real ROAS – not the best-looking surface metrics. Dragalinos Limited treats this integrated measurement layer as the glue that holds the specialist model together operationally.

Why the Model Matters More Than the Metrics

A UA dashboard that looks healthy and a UA program that’s actually working are not always the same thing – and the gap between them is almost always rooted in traffic quality, in measurement pointed at the wrong targets, or both running quietly in parallel. Vanity metrics are good at hiding this. Real ROAS tends to expose it, because the question it asks – are the users being acquired actually doing anything valuable with the product? – It’s harder to dress up when the answer is no.

Dragalinos Limited’s approach is built around not letting the dashboard substitute for the underlying reality. Channel-specialist depth, traffic quality as a hard requirement, and measurement that runs all the way through to revenue – these reinforce each other in ways a generalist setup can’t replicate. Dragalinos Limited has seen this play out in enough different contexts to be fairly confident the structure is what makes the results consistent, not circumstance.