Question 1

What happens in the first thirty days?

Accepted Answer

We start with measurement, not media. The first phase is an audit of what you can currently prove: tracking setup, conversion actions, attribution coverage, and how much of your revenue can actually be traced back to a channel. The typical sequence: - Week one: access, audit, and a baseline of what is measurable today - Week two: fix the tracking gaps that would otherwise corrupt every decision after - Week three: account restructure and budget reallocation based on real outcome data - Week four: first reporting cycle and the ninety day plan We do not make major bid or budget changes until the measurement underneath them is trustworthy.

Question 2

How long before we see results?

Accepted Answer

Measurement improvements show up immediately. Performance improvements typically take sixty to ninety days. The honest reason is sample size. At most budgets, a thirty day window produces only a handful of conversions per keyword or ad group, which is not enough to make a confident decision. We generally require sixty days of data before pausing or reallocating anything material, because decisions made on two or three conversions are noise, not strategy. Anything that promises a transformation in two weeks is either optimizing a channel that was badly broken, or is not measuring carefully.

Question 3

Who will I actually work with?

Accepted Answer

You work directly with the people doing the work. Engagements are led by a senior strategist who stays on the account for its lifetime, supported by specialists in paid media, measurement engineering, creative, and lifecycle. There is no account manager layer translating between you and the practitioners. The person who builds your attribution model is the person who explains it to you.

Question 4

Do you work with in-house marketing teams?

Accepted Answer

Yes, and it is often the best structure. We commonly operate in three modes: - Full service, where we own strategy and execution end to end - Embedded, where we run specific channels alongside your in-house team - Infrastructure, where we build the measurement and data layer and your team runs media on top of it The third is increasingly common. Many teams can execute well but cannot prove what is working, and that is a build problem rather than a media problem.

Question 5

What industries do you work in?

Accepted Answer

Our work concentrates in considered purchase categories where a lead is not the same thing as a sale. That includes logistics and transport, financial services, home services, healthcare, and B2B services. The common thread is a gap between the form submission and the revenue. If your business closes instantly at checkout, a lot of what we build is unnecessary. If a human touches the deal between the click and the cash, the measurement problem we solve is probably your biggest one.

Question 6

How do you decide where to allocate budget?

Accepted Answer

By outcome, not by lead volume. Every keyword is scored on what happened after the click, using data pulled back from your CRM. We build a single table per keyword: clicks, spend, leads, cost per lead, closed deals, revenue, cost per closed deal, and the distribution of CRM outcomes on the leads it produced. Keywords then fall into four groups: - Real volume and real close rate, which get funded and checked for budget caps - High volume and poor close rate, which is where most budget leaks - Low volume and strong close rate, which are usually underfunded - Low volume and poor close rate, which get cut once the sample is large enough to justify it Without the CRM outcome data, the second group looks identical to the first.

Question 7

Why are my highest traffic keywords not producing revenue?

Accepted Answer

Usually because they carry research intent rather than buying intent. Terms built around cost, price, per mile, how much, and estimate attract people who are still learning what something should cost, not people ready to transact. In one account we analyzed, the single most clicked keyword produced zero leads across a seventy four day window. Grouping all research intent terms together, they accounted for roughly a third of total clicks and a fifth of total spend, and produced no closed business at all. This is not a reason to ignore those terms. The intent is real, it is just early. Those keywords are usually a content and organic opportunity rather than a paid one.

Question 8

Should we use value-based bidding?

Accepted Answer

Only if the value you report is the value you actually collect. This is a common and expensive trap. Many businesses report the quoted or estimated price as the conversion value, because that is the number available at form submission. If your cheaper quotes close at a higher rate than your expensive ones, value based bidding will systematically push budget toward the expensive leads that never close. We check the relationship between quoted value and collected revenue before ever enabling it. If they diverge, value based bidding is not merely imprecise, it actively works against margin.

Question 9

How do you handle Performance Max?

Accepted Answer

Carefully, and with reporting built outside of the platform. Performance Max works well when the conversion signal feeding it is clean and when the asset group is genuinely well built. It fails quietly when it is allowed to absorb brand traffic or when it optimizes toward a conversion action that does not represent real business value. Our standard approach is to instrument first so we can see what Performance Max is actually claiming credit for, separate that from what it incrementally caused, and constrain it with exclusions and signals rather than trusting the default configuration.

Question 10

Do you manage Microsoft Ads and other search engines?

Accepted Answer

Yes. Microsoft Ads is frequently underweighted, and in several categories it delivers a materially lower cost per acquisition because competition is thinner. We treat it as its own channel rather than a copy of the Google account. Audience composition, device mix, and match type behavior differ enough that a straight import usually underperforms what the channel can do.

Question 11

What is multi-touch attribution and do I need it?

Accepted Answer

Multi-touch attribution distributes credit for a conversion across the touchpoints that preceded it, rather than giving all of it to the last click. You need it when your buying cycle involves more than one interaction and more than one channel. You do not need it when you run a single channel with a same session purchase, where last click is already an accurate description of reality. The more important question is usually not which model to use. It is whether the outcome data feeding any model is accurate in the first place. A sophisticated model built on unvalidated conversions produces confident, wrong answers faster than a simple one.

Question 12

What is media mix modeling and how is it different?

Accepted Answer

Media mix modeling, sometimes called marketing mix modeling or MMM, uses aggregate time series data to estimate how much each channel contributed to overall business outcomes. It does not require user level tracking, which is why it is returning to favor as click level identity degrades. The difference in one line: multi-touch attribution follows individual journeys, media mix modeling infers channel effects from patterns in aggregate spend and revenue over time. They answer different questions and work best together. Attribution tells you which keyword to pause this week. Media mix modeling tells you whether the channel deserves its budget this quarter. Both consume the same underlying requirement, which is a clean, validated record of actual business outcomes.

Question 13

Why does my ad platform report more conversions than my CRM?

Accepted Answer

This is the most common measurement discrepancy we see, and there are usually several causes stacked together. The frequent ones: - The platform counts form submissions, while your CRM counts qualified opportunities. These are different events and the gap between them is your qualification rate - Double counting, where the same outcome is recorded by both a page tag and an offline import under different identifiers - View-through and modeled conversions, which the platform includes by default and your CRM has no concept of - Different attribution windows, where the platform credits a conversion to the day of the click and your CRM records the day of the sale The fix is not to force the numbers to match. It is to decide which system is the source of truth for revenue, and then make the ad platform learn from that system.

Question 14

What is a CRM feedback loop and why does it matter?

Accepted Answer

A feedback loop sends the real outcome of each lead back to the advertising platform, so that bidding algorithms optimize toward revenue instead of toward form fills. It works in three parts: 1. Capture an identity and a click identifier when the lead arrives 2. Let the CRM record what actually happened, whether that is closed, disqualified, unreachable, or out of service area 3. Send that verdict back to the platform, including retracting leads that turned out to be spam or duplicates Most accounts implement the first step and stop. The result is an algorithm confidently optimizing toward the wrong thing, because nobody ever told it which leads were worth having.

Question 15

How much of my traffic can you actually attribute?

Accepted Answer

It varies by account, and we measure it rather than estimate it. As a reference point, in a recent audit only about thirty percent of leads carried a Google click identifier, while over seventy percent carried an email address. That gap is typical and it has a strategic implication: for many businesses the most reliable identity key is no longer the click ID, it is the contact record. We report attribution coverage as a first class metric, broken out by how each conversion was resolved. When the unresolved share grows, something upstream has broken, and you want that visible rather than averaged away.

Question 16

Can you fix attribution without adding more tracking scripts?

Accepted Answer

Often, yes. Most attribution problems we find are not missing tags. They are configuration errors, broken parameter handling, or outcome data that exists in the CRM and never travels back to the platform. Adding another script to a site that already has a measurement problem usually makes the problem harder to diagnose. We audit what is already firing before proposing anything new, and a meaningful share of engagements end with fewer scripts running than when we started.

Question 17

What is incrementality testing?

Accepted Answer

Incrementality testing measures what would have happened if you had not run the ad. It is the only method that answers the causal question directly, rather than inferring it from correlation. The usual approaches are geographic holdouts, where you suppress a channel in matched regions and compare, or audience splits where a randomized group is excluded. Brand search is the classic case. It almost always looks spectacular in an attribution report and frequently shows far less incremental lift under a holdout, because many of those people were going to find you anyway. We recommend testing on any channel taking a large enough budget share that being wrong about it would matter.

Question 18

What does AI SEO actually mean?

Accepted Answer

Two related things. First, optimizing so your business is cited by answer engines and AI assistants, which increasingly sit between a search and a click. Second, using machine assistance to work at a scale manual processes cannot reach, such as clustering large query sets or generating and evaluating page variants. What it does not mean is mass generated content published without review. That approach worked briefly and now reliably underperforms. The durable advantage is structure, genuine subject expertise, and being the source that answer engines find worth quoting.

Question 19

How do you decide what content to build?

Accepted Answer

From paid search data, wherever it exists. Paid search is the fastest content research instrument available, because you can buy a keyword, observe exactly what it produces, and know its commercial value within weeks instead of waiting for a page to rank. That gives us two content sources with very different logic: - Terms that convert well in paid are proven commercial intent you are currently renting. Building organic pages for them lowers blended acquisition cost over time - Terms with heavy traffic and poor conversion are early stage intent. They usually deserve content rather than bids, because the person is researching and will come back The second group is where most of the wasted paid spend lives and where most of the organic upside sits.

Question 20

How long does SEO take to work?

Accepted Answer

For established domains with existing authority, meaningful movement usually appears in three to six months. For newer domains, six to twelve is more realistic. Technical fixes can move things faster, sometimes within weeks, because you are removing an obstacle rather than building authority. Content programs compound and are slow at the start by nature. We report leading indicators well before rankings move, including crawl coverage, impression growth on target clusters, and query level visibility, so progress is visible before revenue arrives.

Question 21

Do you do technical SEO?

Accepted Answer

Yes, and it usually comes first. Content built on a site that cannot be crawled, rendered, or indexed properly is wasted work. The standard pass covers crawl and index coverage, rendering behavior for JavaScript heavy pages, site architecture and internal linking, structured data, Core Web Vitals, and duplicate or parameterized URL handling. On JavaScript frameworks in particular, we check what search engines actually render rather than what the page appears to contain in a browser, because those are regularly not the same thing.

Question 22

How do you test creative?

Accepted Answer

Against outcomes, and with enough volume to mean something. The common failure is declaring a winner from a click through rate difference on a few hundred impressions. Click through rate is a cheap signal and frequently anticorrelated with quality, because the most clickable message is often the one that overpromises. We test on downstream outcomes wherever volume allows, hold variables steady so a result is interpretable, and report which creative produced revenue rather than which produced engagement.

Question 23

Do you build landing pages?

Accepted Answer

Yes, and we treat the landing page as part of the ad rather than a separate asset. A frequent and invisible problem is the same keyword performing very differently depending on which page it lands on. Because landing page assignment often varies day to day within an ad group, that difference gets averaged away in standard reporting and nobody sees it. We report keyword performance split by landing page so the creative effect and the keyword effect can be separated, which is usually the difference between a page problem and a targeting problem.

Question 24

Which CRMs do you integrate with?

Accepted Answer

We work with HubSpot, Salesforce, GoHighLevel, Pipedrive, Zoho, and custom systems built on Postgres or Supabase. The platform matters less than whether it records a usable outcome. What we need is a clear verdict on each lead, a timestamp, and an identifier that can be joined back to the original click or contact. Where that does not exist yet, building it is usually the first project.

Question 25

Can you build automation into a custom CRM?

Accepted Answer

Yes. A significant part of our engineering work is instrumenting custom CRMs so they function as measurement systems rather than only as record keeping. That typically includes automated lead capture and enrichment at the point of contact, follow up sequences across email and SMS, conversational agents that qualify and capture identity before a human is involved, outcome labeling that is structured enough to analyze, and the feedback loop that sends verdicts back to ad platforms. The measurement value is the point. Every conversation an automated agent handles is either an outcome label you captured or one you lost.

Question 26

How fast should we follow up with a lead?

Accepted Answer

Minutes, not hours. Contact rates decline sharply with delay, and the decline is steepest in the first hour. Speed to lead is also a measurable property of the traffic source itself, which most teams never analyze. Some keywords and channels reliably produce leads who answer the phone, and others reliably produce leads who go dark. That difference belongs in your budget decisions, not just in your sales process.

Question 27

What does your reporting look like?

Accepted Answer

A live dashboard rather than a monthly slide deck, plus a written analysis on a regular cadence that explains what changed and what we are doing about it. Every reported figure is labeled as measured or modeled. If a number is an estimate, it says so. If a conclusion rests on a sample too small to support it, that appears alongside the conclusion rather than in a footnote. You also get the underlying data. We do not treat the measurement layer we build as something you rent from us.

Question 28

Who owns the data and accounts?

Accepted Answer

You do, in all cases. Ad accounts, analytics properties, tag containers, dashboards, and any data infrastructure we build are created under your ownership or transferred to it. If an engagement ends, nothing switches off and nothing has to be rebuilt. We consider account ownership leverage to be a red flag in this industry rather than a business model.

Question 29

Why do my dashboard numbers keep changing?

Accepted Answer

Some movement is legitimate and some is a defect, and it is worth separating them. Legitimate causes include conversion lag as delayed conversions get attributed back to their original click date, platform restatements of modeled conversions, and currency or timezone boundaries. Defects include silently failing data loads, partitions that were never registered, and filters that quietly exclude rows. A specific one worth checking in any account: automated health check or monitoring records living in the same table as real records. We have seen cases where those synthetic rows outnumbered real ones by more than forty to one, which corrupts every rate calculated from that table while every individual record looks perfectly valid.

Question 30

How do you price engagements?

Accepted Answer

A monthly retainer based on scope and complexity, not a percentage of ad spend. Percentage of spend creates a direct conflict of interest. It rewards the agency for spending more regardless of whether more spending is the right call, and it penalizes the agency for the efficiency work that often produces the best return. Build projects, such as attribution infrastructure or CRM instrumentation, are usually scoped as fixed fee separate from the ongoing retainer.

Question 31

Is there a minimum ad spend or contract length?

Accepted Answer

There is no hard minimum spend, but below roughly ten thousand dollars per month in media the statistical reality limits what can be proven. At low volume, most performance differences are indistinguishable from noise, and we would rather say that than sell a program that cannot be evaluated. Initial agreements typically run ninety days, which reflects how long honest measurement takes. After that, engagements are month to month. We do not use long lock-ins.

Question 32

Do you take a percentage of ad spend?

Accepted Answer

No. Media budget goes directly to the platforms and you see the invoices. Our fee is a flat retainer for the work, so our recommendation to reduce spend on something that is not working costs us nothing. That alignment matters more than it sounds, because a large share of the value in the first ninety days usually comes from stopping waste rather than adding budget.

How can we help?

Browse by category

WORKING WITH US

PAID SEARCH & PPC

ATTRIBUTION & MEASUREMENT

AI SEO & ORGANIC

PERFORMANCE CREATIVE

EMAIL, CRM & AUTOMATION

REPORTING & ANALYTICS

PRICING & AGREEMENTS

Things that cost real money

The conversion action type trap

Unrecognized tracking macros fail silently

The toggle you see may not be the one bidding reads

Click level data has a hard ninety day wall

Your best identity key may not be the click ID

The most clicked keyword can produce nothing

Still need help?