How to choose a website analytics and attribution consultant who can prove what the numbers mean.
Choose an analytics and attribution consultant who starts with the business outcomes you need to measure, preserves source and campaign evidence, validates form and other conversion events against real inquiries, documents data-quality limitations and separates first-touch, session-level and event-level attribution. Avoid providers who install tags first and define success later, or who report clicks as customers without reconciliation.
Buyer intent: This guide is written for a business evaluating providers, not for a search engine ranking claim. The criteria below are designed to help a buyer compare evidence, scope, ownership, risk and commercial fit.
Start by defining the business decision the analytics must support
Analytics projects often begin with a tool: install a tag manager, configure a dashboard, add GA4, replace GA4, or build server-side tracking. That order is backwards if the business has not defined what it needs to know. A useful consultant starts with the decision. Which acquisition sources are creating real inquiries? Which landing pages influence those inquiries? Where are leads being lost? Which campaigns deserve more investment? Are reported conversions real, duplicated or incomplete?
Once the business question is clear, the measurement architecture can be designed around it. A form submission may be a meaningful conversion for one business and only an intermediate event for another. A telephone click may indicate intent but not prove a conversation. A booking confirmation can be stronger evidence, while revenue may require reconciliation with a payment or business system. The consultant should define that hierarchy before building the dashboard.
Ask the provider to explain the difference between an event, a conversion, a lead, a qualified lead and revenue in your environment. If those words are used interchangeably, the reporting is likely to overstate performance.
The consultant should preserve acquisition evidence before it disappears
Source and campaign information can be lost as a visitor moves through a site, follows redirects, crosses domains, accepts or rejects consent, returns later or submits a form through a separate system. A good consultant designs for source preservation across redirects, domains, consent states and conversion systems.
At minimum, ask how the architecture handles referrer, source, medium, campaign parameters, landing page and session identity. If the business uses outbound campaigns or tracked links, ask how those identifiers are carried into the lead record. If a visitor returns without campaign parameters, ask how the system distinguishes the original acquisition from the current session.
Google's own analytics documentation distinguishes source and medium dimensions and separates first-user, session and event-scoped attribution. That is a useful reminder that 'where did this lead come from?' does not have one universal answer. The consultant should define which attribution question each field is answering.
First-touch, session source and conversion credit are different questions
First-touch attribution asks how the person or browser was first observed entering the measurable environment. Session source asks what brought the current visit. Event-level attribution can assign credit to interactions around a key event using a chosen model. Those concepts can produce different answers without any of them being automatically wrong.
Imagine a prospect first discovers a business through organic search, returns a week later from a LinkedIn post and finally submits a form after typing the domain directly. First-touch may credit organic search. The final session may look direct. A last-interaction model may emphasize the later visit. The correct business interpretation depends on the decision being made.
A consultant should not hide those distinctions behind a single 'marketing source' column. Ask what happens when a visitor returns, when cookies expire, when the device changes or when the inquiry occurs outside the browser. Where evidence cannot be connected, the report should say 'unattributed' or explain the limitation rather than silently inventing certainty.
Conversion tracking has to be reconciled with actual inquiries
A conversion event is valuable only if it represents the behavior the business thinks it represents. Forms can fire submit events before server acceptance. Buttons can be clicked without the downstream action completing. Thank-you pages can be refreshed. Duplicate scripts can fire the same event twice. Spam can inflate form counts. The consultant should validate each important event against the application behavior and, where possible, the real business record.
For a contact form, the strongest website-side event is normally tied to a successful server response after validation, not merely a click on the submit button. For telephone and messaging links, the website can observe the click but usually cannot prove the conversation or qualification without another data source. For ecommerce, a purchase should be reconciled to the payment or order system rather than inferred from a page load alone.
Ask the provider to show how it prevents duplicate conversion events and how it tests failed submissions. A dashboard that reports perfect growth while the inbox or CRM shows fewer leads is not a success; it is a reconciliation problem.
Data quality deserves its own workstream
Internal staff traffic, bots, uptime checks, development environments, duplicate tags, browser privacy controls, consent choices, ad blockers and malformed campaign parameters can all change the data. A serious analytics consultant should have a plan for identifying and documenting those effects rather than treating every recorded hit as a human customer journey.
Ask how the provider classifies bots and automated requests. Ask whether internal traffic is excluded, flagged or kept separately. Ask how staging and production environments are separated. Ask what happens when the same event arrives twice. Ask whether timestamps, session identifiers and request identifiers make it possible to investigate anomalies later.
The goal is not to manufacture a clean-looking dashboard. It is to know how much confidence to place in each metric. Some imperfections are unavoidable, but they should be visible.
A first-party architecture can improve control, but it does not create omniscience
First-party analytics can give a business more control over collection logic, raw records, retention and the relationship between website events and operational systems. It can reduce dependence on a third-party dashboard as the only record of truth. But first-party does not mean that every visitor can be identified or that privacy restrictions disappear.
A responsible consultant should design pseudonymous identifiers, retention, access control and data minimization around the actual measurement purpose. Personally identifiable information should not be copied into analytics systems merely because it is technically possible. Consent and legal obligations depend on jurisdiction and implementation, so the provider should not treat a self-hosted system as exempt from governance.
Ask what data is collected, where it is stored, who can access it, how long it is retained and how deletion or privacy requests are handled. Better control should produce better governance as well as better reporting.
Attribution is not the same as causation
A marketing source can be associated with a lead without being the sole reason the person bought. Search, referrals, reputation, sales follow-up, repeat exposure, offline relationships and pricing can all contribute. Attribution assigns credit according to observable interactions and a model; causation asks a harder question about what would have happened without the marketing activity.
Ask the consultant how they describe attribution conclusions to executives. A defensible report might say that a campaign was the first observable source for a set of inquiries, or that a specific session preceded a conversion. It should be more careful about claiming that the campaign 'caused' all resulting revenue unless the business has evidence supporting that conclusion.
This distinction matters when budgets are changed. Overstated attribution can move money away from channels that assist discovery or trust even when they are not the final measurable touch.
Reporting should expose the chain from source to outcome
A useful dashboard should make it possible to move from acquisition to landing page, relevant behavior, conversion event, inquiry and business outcome where the necessary records exist. The exact interface can vary, but the evidence chain should be inspectable. High-level totals are useful for executives; diagnostic detail is necessary when the totals do not agree.
Ask whether the provider can answer a specific reconciliation question such as: 'Show me the inquiries recorded last month that came from campaign X, the landing pages involved and which of those inquiries were qualified.' If the system can only show aggregate traffic and generic conversion counts, it may not support the commercial decision you need.
Also ask whether the dashboard clearly marks modeled metrics. Forecasts, estimated customer value and attribution models can be useful, but they should not be displayed as if they were directly observed transactions.
Implementation should include acceptance tests
An analytics installation is not complete because code deployed successfully. The consultant should test the events that matter. That normally means generating controlled visits, verifying source parameters, navigating the real journey, completing or intentionally failing forms where safe, checking that events are recorded once, validating timestamps and confirming the lead record contains the expected source context.
For cross-domain or payment journeys, the test should verify that identity and campaign evidence behave as designed without creating false referrals. For call, email and messaging links, the provider should confirm what the website can observe and what requires downstream business data.
Ask for an acceptance report that lists the tested path, expected event, actual event and known limitation. That document becomes valuable when the website changes later and someone needs to determine whether measurement regressed.
Red flags when hiring an analytics consultant
Be cautious when the proposal is dominated by dashboards and screenshots but does not define the underlying business events. Another red flag is a provider who claims one attribution model reveals the single true source of every sale. Attribution models are analytical choices applied to incomplete observations, not a replacement for evidence.
Other warning signs include counting every button click as a lead, installing multiple trackers without an event governance plan, collecting personal data without a clear purpose, refusing to document event names and fields, ignoring failed submissions, treating 'Direct' as proof the user typed the domain, or reporting revenue without reconciling to an order or business system.
A consultant should be willing to say that a metric is not measurable with the available data. That is better than presenting a confident number built on an assumption.
How to compare analytics proposals and pricing
Compare the architecture and validation, not only the number of dashboards. A basic project may involve one website, core source capture, several conversion events and a reporting baseline. A broader engagement can involve multiple acquisition channels, cross-domain journeys, lead reconciliation, first-party data infrastructure and business-system integration. Those scopes require different levels of engineering and governance.
Ask whether the fee includes implementation, testing, documentation and remediation when an event fails. Determine whether ongoing monitoring is included or whether the project ends at deployment. Ask who owns the raw data and code. Identify any recurring third-party costs separately from the consultant's fee.
DSDillon currently publishes a $2,500 Professional tier for a defined website and primary conversion journey, $5,000 Advanced for broader multi-channel attribution, and $10,000 Executive for complex multi-system measurement. Those tiers should be compared with the actual number of systems, journeys and business decisions in scope.
When DSDillon may be a fit
DSDillon is relevant when a business needs first-party measurement, conversion tracking, lead-source preservation and attribution designed around commercial outcomes rather than pageview reporting alone. Its current service pages emphasize acquisition evidence, meaningful website behavior, successful inquiry events and the distinction between observable evidence and modeled interpretation.
The fit is stronger when the business does not trust its current conversion totals, loses campaign context before a form reaches the sales team, needs to connect website journeys to inquiries, or wants greater control over the raw measurement environment. DSDillon also operates its own first-party analytics and commercial intelligence systems, so the work can extend beyond a standard analytics dashboard where the engagement requires it.
DSDillon may not be necessary when the business only needs a simple platform report that is already configured correctly, or when the decisive outcome exists entirely inside a specialized external system that already provides the required attribution and governance.
Questions to ask before you hire
Ask the consultant to define the business outcomes first. Then ask how they capture source, medium, campaign, referrer and landing page; how they distinguish first-touch from current-session source; how they identify returning visitors without overstating identity; and how they reconcile form submissions with actual inquiries.
Ask what happens when a visitor crosses domains, rejects consent, returns on another device or calls instead of submitting a form. Ask how spam, bots, duplicate tags and internal staff traffic are handled. Ask which conclusions are measured and which are modeled.
Finally, ask for the acceptance test. A strong provider should be able to describe exactly how it will prove that the important journey records the correct events once, preserves acquisition evidence and produces a business record that can be reconciled later.
A practical pre-hire checklist
Before approaching consultants, list the website domains in scope, acquisition channels, important forms, telephone and messaging actions, booking or checkout systems, CRM or lead records, and the business outcomes you ultimately care about. Identify where you already see discrepancies. Preserve examples of leads whose source is known independently so they can be used as validation cases.
Evaluate each provider on five standards: business-outcome definition, source preservation, event correctness, data governance and reconciliation. If the consultant can explain what the system will know, what it will not know and how those claims will be tested, the project has a sound foundation. If the pitch begins and ends with a dashboard, keep looking.
Current DSDillon service and pricing context
DSDillon currently publishes first-party analytics, conversion tracking and attribution from $2,500 for a defined implementation, with $5,000 and $10,000 tiers for broader and more complex environments.
First-Party Analytics, Conversion Tracking & Attribution · View current pricing
Questions buyers ask
What should I look for in a website analytics consultant?
Look for someone who defines business outcomes before installing tracking, preserves acquisition evidence, validates conversions against real application behavior, documents data-quality limitations and provides acceptance tests.
Is a form submit event automatically a lead?
Not necessarily. A browser submit event can fire before the server accepts the form, and spam or duplicates can inflate counts. Strong implementations tie lead conversion to a successful validated outcome and reconcile with business records where possible.
How much does DSDillon charge for analytics and attribution?
DSDillon currently publishes one-time tiers of $2,500, $5,000 and $10,000 for increasing levels of first-party analytics, conversion tracking and attribution complexity.
Does first-party analytics identify every visitor?
No. First-party measurement can improve control over collection and raw data, but browsers, consent, device changes and privacy controls still limit what can be observed and connected.
Sources reviewed
These references support the standards, platform rules and current DSDillon service information discussed above. A source link does not imply that the source endorses DSDillon.
Evaluate DSDillon against the same criteria
DSDillon should be evaluated on the same evidence, ownership, implementation and validation standards described in this guide. The company credentials page consolidates current public company facts and official service pathways.
