Why a web-only plan stops short for leads
The generic plan treats an event on the site as proof that a goal was met. For an enquiry-led business that event is the start of the process, not the end of it. A completed form says a person typed their details; it says nothing about whether they own a roof, have a budget or ever answer the phone. Two campaigns can deliver the same number of submissions and very different revenue, and a plan that reports submissions will reward whichever one produces the cheapest enquiries.
The data is also split across three systems with three clocks. Analytics records the session and the campaign, the CRM records what sales did with the lead, and the ad platform records the click and its cost. None of them can answer the objective alone. Finally, the conversion the plan names is not only a reporting choice: ad platforms bid towards whatever they are told counts, so the definition doubles as an instruction to the algorithm. Which goals belong in the plan at all depends on whether the programme captures existing demand or creates it, the distinction drawn in the lead generation versus demand generation comparison.
Decisions that change for lead generation
| Plan component | In the generic plan | In a lead-generation plan | Decision it forces |
|---|---|---|---|
| Goals and KPIs | One observable goal per objective, proved on the site | A ladder of goals from enquiry to contract | How many depths, and which one the business is judged on |
| Tracking specification | Events and parameters in analytics | Events plus hidden fields, CRM properties and an identifier | Where the lead identifier is minted and who keeps it intact |
| Source of truth | The analytics property | Analytics for sessions, CRM for stages and money | Which system wins when the counts disagree |
| Feedback to media | Not covered | Qualified outcomes imported into ad platforms | How often, at which stage, with which value |
| Capture surfaces | The website | Website forms, calls and native platform forms | How off-site leads join the same spec |
| Consent | Mentioned as a data limit | Recorded per lead and passed to every system that uses it | What happens to a record without consent for advertising use |
Define conversions at several depths
Write a separate goal for each stage the business recognises: the enquiry, the qualified lead, the opportunity and the won deal. Each gets its own definition, its own owner and its own KPIA metric tied directly to an objective., and the definitions come from sales, not marketing, because sales decides what qualified means. The trade-off is between speed and truth. Shallow stages arrive quickly and in volume but say little about value; deep stages are reliable but slow and sparse. Choose the qualified lead as the headline when the sales cycle is long, and the opportunity when sales works leads within days. If a scoring model sets the qualified threshold, the lead scoring guide is where that threshold is built and calibrated; the plan only references it and dates any change.
Map CRM stages with a shared lead identifier
Every depth beyond the form lives in the CRM, so the plan must say how a CRM stage is joined back to the session and the click. That means a lead identifier created at capture and written to the analytics event, the hidden form fields and the CRM record together, plus the click identifiers and campaign tags (UTMURL tags identifying traffic source. values) stored on the record. The alternative, matching on email address later, works for reporting but fails for anything that needs the click. Choose identifier-first whenever paid media is more than a small part of the mix. The guide to connecting the CRM to ads and analytics covers the capture mechanics; the plan's job is to name the fields, who owns them and which stage changes trigger an export. Phone and chat enquiries need the same identifier, written by the call trackingAttributing phone calls to the marketing source that produced them, usually by showing different tracking numbers per source or visitor. or chat system, and the guide to measuring phone calls and chat enquiries covers how they are counted, attributed and joined to the CRM.
Schedule the offline conversion import
Once stages can be joined to clicks, they can be sent back. The plan records which stages are imported, how often, with what value, and which one is primary for bidding. A frequent import helps the platform learn but multiplies the chance of duplicates and late corrections; a slow one is tidy but leaves bidding blind for days. Choose a daily import of the qualified stage as the default, keep deeper stages as secondary conversions, and set values from expected revenue per stage rather than a flat number. The conversion tracking set-up guide covers the website side of the same actions, which must use matching names so the two do not double count.
Put native platform lead forms in the same spec
Lead forms inside social platforms never touch the website, so they produce no analytics event at all. Leaving them out makes the plan describe only part of the funnel; treating them as a separate world makes cost per qualified lead impossible to compare. Specify them as a capture surface with the same identifier, the same required fields and the same qualification definition, synced to the CRM by a named connector. The LinkedIn targeting and lead gen forms guide explains the form fields and routing; the plan adds the source value and the stage mapping. Choose fewer native-form fields only when the qualification questions can still be answered later by sales.
Capture consent at the form and carry it downstream
A lead record moves into email, the CRM, enrichment tools and audience uploads, and each use may need its own permission. The plan records what consent is collected at each capture surface, where it is stored and which downstream uses it allows. The tension is between completeness and compliance: excluding leads without advertising consent from imports makes the platform's view smaller but defensible. Choose the defensible version and note the gap as a known limitation.
Decision points in a worked lead-generation plan
A measurement plan for an invented solar installer, Brightrow Solar
Brightrow Solar installs rooftop panels across two regions. It spends £38,000 a month on search, social and native lead forms, receives about 620 enquiries a month and signs roughly 41 contracts worth £9,200 on average. The sales cycle from enquiry to signature is about six weeks, including a home survey. The plan, condensed:
| Depth | Definition owned by sales | System of record | How it is captured | Ad-platform use | Read |
|---|---|---|---|---|---|
| Enquiry | Valid form, call over 60 seconds or native lead | Analytics and CRM | Event generate_lead with lead_id; call tracking | Secondary conversion | Weekly |
| Qualified lead | Homeowner, suitable roof, survey booked | CRM | Stage change exports lead_id, click ID, value £600 | Primary bidding conversion, imported daily | Weekly, two-week lag |
| Opportunity | Survey done, quote issued | CRM | Stage change, value £3,000 | Secondary conversion | Monthly |
| Won deal | Contract signed, deposit paid | CRM and finance | Stage change, actual contract value | Secondary, value-based reporting | Monthly, six-week lag |
Targets: enquiry-to-qualified rate from 31% to 36%, cost per qualified lead under £190 against a £205
baseline, won deals at 45 a month by the end of the second quarter. Consent: the advertising-use checkbox is
present on all forms; about 12% of leads decline it and are excluded from imports, noted as a limitation.
The native lead forms share the same lead_id format and push to the CRM through a connector owned by the
marketing operations lead.
The plan surfaced one decision before launch: a social campaign producing 40% of enquiries had a qualified rate of 14%, against 38% for search. Bidding on enquiries would have fed it; bidding on qualified leads moves budget away from it without anyone arguing about it.
The value of writing the plan down showed up before a single tag changed. The ladder of depths exposed that two channels with similar enquiry volumes were worth very different amounts, and the identifier decision showed that phone enquiries had no way into the CRM stage export until call tracking wrote the same field.
Measuring the plan: KPIs and review rhythm
The headline measures for a lead-generation plan are cost per qualified lead, the conversion rate between each pair of stages, pipeline value and won revenue per channel, and the time each stage takes. The CPA calculator is useful for setting the cost ceiling per stage once the stage rates are known. Alongside these, the plan needs health measures of its own: the share of CRM records that carry a lead identifier and a click identifier, the match rate reported by each import, and the share of leads missing a source. When those fall, every channel number above them becomes less trustworthy, so they belong at the top of the weekly review rather than in an appendix.
Cadence follows the lag of each depth. Enquiries and import health are read weekly. Qualified rates are read weekly but only for leads old enough to have been worked. Opportunity and won-deal measures are read monthly and always by enquiry cohort rather than by the month the deal closed, so a campaign is judged on the leads it produced. The AttributionAssigning conversion credit to touchpoints. setting used for each imported conversion is recorded and dated, as is the GA4Google's event-based analytics platform. key-event list, so a later change is visible. The tracking plan holds the field-level detail beneath all of this.
Where a lead-generation plan misleads
Imported stages describe only the leads that could be matched, and the unmatched share is rarely random: phone leads, leads without consent and leads from devices that blocked identifiers are under-represented. Stage definitions drift when sales changes process, and a qualified rate can jump because the bar fell rather than because marketing improved. Long cycles mean the most recent month always looks worst. And the plan measures enquiries the business can see; referrals and offline word of mouth still produce contracts that the chain credits to whatever click came last.
Common mistakes
- Stopping at the thank-you page. A plan that ends at the form judges channels on volume, not value.
- Letting marketing define qualified. The definition has to come from the people who work the leads.
- Minting the identifier in the CRM. By then the session and click are gone and cannot be joined.
- Importing every stage as primary. Bidding needs one target; the rest are context.
- Forgetting native forms and calls. Leads that skip the website still need the same fields and stages.
- Reading deep stages by close month. Cohort by enquiry date or every recent campaign looks like a failure.
