Yes. A window and door company should measure marketing past website traffic and form fills, but only after it agrees on what a real sales opportunity and a real sale mean. The useful question is not, “How many leads did we get?” It is, “Which sources brought homeowners we could serve, who took the next step, and who became completed work?”
This article covers a practical system for replacement-window and door contractors that generate inquiries online, by phone, through a showroom, or through referral partners. Traffic shows whether a campaign is being seen. Raw leads show whether people respond. Neither tells you whether marketing is producing workable opportunities or closed jobs.
The goal is a clean handoff from an ad, search listing, referral, or mail piece into your sales process, then a record of what happened next. You need shared definitions, consistent source information, and a steady review rhythm.
Start with the outcome you are trying to manage
A homeowner who asks for a brochure, a caller who wants a repair you do not offer, and a homeowner who sets an in-home consultation are not the same result. Counting them equally makes weak marketing look strong and can hide a sales-process problem.
Build the plan around the job your team actually sells. A replacement contractor may use inquiry, qualified opportunity, appointment set, appointment held, proposal, signed agreement, and completed installation. A showroom-led business may add a showroom visit. Use the path that reflects your operation, not a generic funnel copied from software.
Marketing and sales should agree on the stages before anyone debates channel performance. That agreement is the backbone of a window and door marketing strategy built around pipeline data. It lets the owner see whether a problem starts with targeting, follow-up, scheduling, pricing, financing, or the mix of work being pursued.
Define stages so they can be used the same way every time
Keep stage definitions observable. Avoid labels such as “good lead” or “hot prospect” unless the team can say what qualifies someone for that label. A salesperson should not need to guess which status to choose after a rushed phone call.
| Stage | Operational definition | Who updates it | What to record |
|---|---|---|---|
| Inquiry | A person asks for information, a quote, a consultation, or calls with a sales-related need. | System or intake staff | Date, contact method, requested product, source information |
| Qualified opportunity | The project fits your service area and work type, and the customer is a plausible fit for your sales process. | Intake staff or salesperson | Qualification result and a standardized reason when not qualified |
| Appointment set | A specific showroom or in-home meeting is on the calendar. | Scheduler or sales team | Appointment date, meeting type, assigned rep |
| Appointment held | The meeting took place. | Salesperson | Held or no-show status and next action |
| Proposal issued | A priced proposal or scope was delivered. | Salesperson | Proposal date, product category, amount in the CRM or estimating system |
| Sold | Your company has a defined commercial commitment, such as a signed agreement or paid deposit. | Sales or office | Sold date and contract amount |
| Completed or canceled | The work is completed, or the sale later falls apart under a defined reason. | Operations or office | Completion or cancellation date and reason |
A stage is a record of a business event, not a grade on a person. If a salesperson marks an inquiry “unqualified,” require a standard reason, such as outside service area, repair-only request, product mismatch, duplicate, or unable to contact. A blank reason turns a real operating issue into an argument.
Set one definition for “sold.” It may be signed agreement or paid deposit, depending on how your business treats a committed job. Do not allow individual reps or campaign reports to use different definitions.
Keep lead quality separate from lead volume
A weekly report should be able to answer two different questions:
- Did this source produce enough inquiries to deserve attention?
- Did those inquiries become qualified opportunities, held appointments, proposals, and sales at an acceptable rate for our market and offer?
That separation protects you from two common mistakes. The first is shutting off a source because it generates fewer inquiries even though its opportunities move well through the pipeline. The second is raising spend on a channel because its form-fill count looks strong while the sales team rarely reaches those people or their projects do not fit.
Do not use one universal “cost per lead” target. A brand search, referral, paid social offer, neighborhood mailer, and showroom walk-in enter at different points in the decision process. Compare like with like, using cost per qualified opportunity, held consultation, and sale within a source or campaign.
Capture the source before it disappears
Source data is most useful when it is captured at intake and carried forward with the record. Trying to reconstruct source months later from a salesperson’s memory or an invoice note usually produces guesswork.
Use both technical and human capture. Technical capture preserves what brought someone to a landing page. Human capture catches calls, showroom walk-ins, word of mouth, and other offline paths.
Create one source taxonomy for the whole company
Pick a short list of source categories and do not let every employee create a new version. The source field should answer where the inquiry came from. The channel field can describe how it got there. The campaign field should identify the specific effort. Keep the values lower case and consistent.
| Field | Example values | Use it to answer |
|---|---|---|
| Source | google, google_business_profile, facebook, referral, yard_sign, showroom | Where did the inquiry originate? |
| Channel | organic_search, paid_search, paid_social, direct_mail, referral | What type of marketing or relationship produced it? |
| Campaign | replacement_windows_fall, entry_doors_neighborhood_a | Which effort should be evaluated? |
| Offer or product | windows, patio_doors, entry_doors, mixed_project | What was the initial interest? |
| Market area | north_county, city_center, west_suburbs | Where is demand arriving? |
This is not busywork. “Google,” “google ads,” “web,” and “internet” in four systems will not produce a dependable report. Put approved values in the CRM, call-intake form, lead-import process, and report. Leave an “other, describe” option, then review recurring exceptions and add them when needed.
For traffic that comes through a tagged link, use consistent UTM values. Google Analytics describes utm_source, utm_medium, and utm_campaign as the core campaign parameters and warns that case differences create separate values. That is why Facebook, facebook, and fb should not all be used for the same platform. Build the campaign URL once, save it in a shared sheet, and use that approved version in ads, email, QR codes, partner links, and printed pieces.
For a direct-mail piece, use a dedicated landing page or tagged QR code. A campaign-specific phone number may fit your operation. For referrals, ask “Who referred you?” and record the person or partner in the CRM, not web analytics.
Source tagging should support the rest of your demand plan, not replace it. If your team is still deciding which local channels, pages, offers, and proof points belong together, use a full demand-generation plan for a window and door company to work through that broader decision first.
Protect original source and latest source
A homeowner may first discover your company through a paid search ad, come back through an organic search result, then call after seeing a neighbor’s installation sign. All three facts can be useful. They are not interchangeable.
Keep an original source that does not change after the first recorded inquiry. Add latest source only if it is reliable and useful. This supports first-touch analysis without pretending one event explains a long buying decision.
For each inquiry, preserve the landing page or call source, source, medium, campaign, date, and a non-identifying record key in the CRM. If staff correct a source, require a reason and retain the original value.
Make offline sales outcomes part of the marketing record
Window and door sales usually close after the web visit: calls, qualification, measurements, meetings, product choices, proposals, contracts, deposits, and installation. If those outcomes stay in sales notes or accounting, marketing reports stop at the least useful point.
Connect your CRM or estimating system to the original lead. Each record needs a stable internal ID, source fields, lead date, current stage, stage dates, proposal and sold amount when applicable, and a lost or canceled reason. The same record should travel from intake through installation.
A weekly export can work when systems do not integrate. Match leads by internal ID, add source data from intake, and reconcile sales and cancellations with office records at month-end. The discipline matters more than early automation.
Google Ads supports measurement of later offline outcomes from ad clicks or calls. Its documentation describes saving a Google Click ID with lead information and later returning it with a conversion type and time. Google also points advertisers toward enhanced conversions for leads. This is a platform-specific project, not a CRM replacement. It needs proper tagging, data handling, consent where required, and testing before it guides budget decisions.
Choose only meaningful milestones, such as qualified opportunity, appointment held, sold, and perhaps completed installation. Keep the platform conversion name tied to the internal stage definition.
Review losses as carefully as wins
A channel is not automatically bad because it produces losses. Loss reasons show what needs attention: repair requests when you sell replacement work, homes outside your service range, or appointments set but not held require different fixes.
Use controlled lost-reason options such as:
- outside service area
- repair or service request only
- product or project mismatch
- duplicate inquiry
- unable to reach after defined attempts
- appointment not held
- selected another contractor
- timing postponed
- financing or budget issue
- canceled after sale
Do not turn this list into opinions about the homeowner. It should show patterns in fit, offer, response, scheduling, pricing, or operations. Review records when a reason suddenly rises.
Track the time from inquiry creation to first attempted contact, then whether contact was made. The stages work best alongside the response workflow that keeps new inquiries from sitting untouched, because a source report means little if calls and forms are not handled consistently.
Use GA4 events as a reporting aid, not as a sales system
GA4 can show website activity, but it does not replace intake, CRM ownership, or sales-stage discipline. Google’s GA4 recommended lead-event guidance lists generate_lead, qualify_lead, disqualify_lead, working_lead, close_convert_lead, and close_unconvert_lead for lead-generation funnels that include online and offline activity.
The names give reporting a common vocabulary. They are not sent automatically, and their use does not guarantee that a form, CRM, call platform, or ad account is connected correctly. Implementation still needs clear triggers, prescribed details where applicable, record matching, and verification.
A practical approach is to map only the events you can define cleanly:
| Recommended event | Possible contractor trigger | Guardrail |
|---|---|---|
generate_lead |
A website form is successfully accepted or an offline inquiry is created. | Do not fire it just because someone opens a form or clicks a phone number. |
qualify_lead |
The CRM record meets the agreed qualification rule. | Use the same definition in sales reports. |
working_lead |
A rep makes meaningful two-way contact or begins active follow-up under your rule. | Define what counts before reporting it. |
close_convert_lead |
The record reaches your one definition of sold. | Do not use it for a proposal merely sent. |
close_unconvert_lead |
The record is closed without becoming a sale. | Preserve a controlled loss reason in the CRM. |
Document each event’s trigger, source system, owner, and test method in a one-page plan. Test with a controlled submission before using the report. GA4 DebugView and Realtime can help confirm event arrival, not whether salespeople selected the right CRM stage or every record matched.
Do not force every stage update into GA4. The CRM or estimating system is the system of record. A small, verified event set is better than a long list nobody can explain.
Keep analytics useful without putting customer details into it
Analytics should show campaign and funnel patterns without becoming a shadow customer database. Keep the CRM for contact and project information. Keep web analytics to aggregated, non-identifying information such as event names, page paths, source, medium, campaign, device category, broad product interest, and non-identifying stage labels.
Google identifies email addresses, mailing addresses, phone numbers, precise locations, full names, and usernames as personally identifiable information in its guidance. Do not send those values to GA4 in event parameters, user properties, URLs, form labels, search terms, or custom dimensions. A form confirmation URL that includes an email address is still a problem if analytics receives the page URL.
Avoid project notes, street addresses, full ZIP codes, free-form responses, or quote amounts tied to an identifiable contact in web analytics. Your privacy notice, consent practices, retention settings, access controls, and vendor setup should fit where you operate and the tools you use.
Use internal lead IDs only in secured systems that need record matching. Give a marketing vendor grouped results such as source, campaign, stage count, and revenue total. Give people only the access required for their job.
Before launching a new form or campaign, run this check:
- Confirm the landing page and thank-you flow do not place names, emails, phone numbers, or addresses in URLs.
- Inspect analytics event parameters and tag-manager variables for form fields that may be passed through by mistake.
- Make sure campaign names describe the campaign, not an individual homeowner or address.
- Test the form, call routing, CRM creation, and stage update using a controlled record.
- Decide who can view raw lead records and who only needs aggregate reporting.
Set a reporting rhythm the team can keep
A good measurement system has different meetings for different decisions. Daily reporting should help the team respond to active inquiries. Weekly reporting should show whether the pipeline is moving. Monthly reporting should guide channel and budget decisions. Trying to make one dashboard serve all three usually creates clutter.
| Cadence | Primary audience | Review | Decision to make |
|---|---|---|---|
| Daily | Intake, scheduling, sales | New inquiries, uncontacted records, appointment changes, broken tracking alerts | Who needs contact or correction today? |
| Weekly | Sales manager and marketing owner | Inquiries by source, qualified opportunities, appointments set and held, proposals, response time, loss reasons | Where is the process leaking or improving? |
| Monthly | Owner and leaders | Spend, source-to-stage conversion, sold work, canceled work, pipeline value, closed revenue when available | What should continue, change, pause, or be investigated? |
| Quarterly | Leadership | Stage definitions, source taxonomy, reporting gaps, market coverage, operational capacity | Are we measuring the business we are actually running? |
The weekly meeting does not need a presentation. Start with exceptions: records with no source, records without timely contact, and sources with sharp changes in qualification, held appointments, or loss reasons. Open a small sample. The report points to where people should look.
At the monthly level, use cohorts. A lead created this month may not be sold this month. Follow January inquiries through later stages as they mature instead of casually crediting current-month sales to current-month spend.
For a group of operating companies, portfolio reporting for multi-location window and door platforms should preserve those same stages while keeping each company’s service area, sales model, and capacity visible. A rollup that removes the local operating context is not a decision tool.
Include capacity. If qualified leads rise but the calendar is full, the decision may concern scheduling, sales coverage, or installation capacity rather than more volume. If appointments fall, check no-shows, confirmation, and time-to-contact before blaming the ad channel.
Build the first version before trying to automate everything
Start with one report containing source, inquiry count, qualified opportunities, appointments set and held, proposals, sold count and value, loss reasons, and stage dates. Reconcile it manually for a few cycles. Fix definitions, missing sources, and duplicates before commissioning a complex dashboard.
When you consider outside help, judge the measurement process as carefully as the channel plan. A comparison guide to window and door marketing agencies can help you frame questions about CRM ownership, source capture, reporting access, stage definitions, and what happens when the data does not line up. You should retain access to the accounts and data that describe your own pipeline.
The standard is straightforward: every important inquiry should be attributable to a source as accurately as practical, moved through a defined sales path, and visible in a report soon enough for someone to act. Traffic and raw leads can stay on the dashboard. They just should not be the final score.