The B2B SaaS Guide to Lead Generation

The six things that make it hard

B2B software runs into the same six obstacles whatever it sells. They are worth naming, because each chapter is a response to one of them.

1. The buyer is a committee, not a person. A hospital administrator, an IT lead and a procurement manager evaluate the same product on different criteria, in a different order, and the advertisement that works for one is noise to the others. A funnel built per channel treats them as one buyer. A funnel built per vertical and persona does not.
2. The conversion you can measure is not the one that pays. A trial, a demo request or a whitepaper download is a moment in a cycle, not revenue. Optimising to it without a view of what happens next produces cheap leads nobody can close.
3. The cycle is long and the numbers arrive late. A practice replacing its clinical software routes capital approval through a healthcare group. A channel partner takes months to transact. Media decisions made on last week’s leads are made on the wrong evidence.
4. The category is crowded by budgets bigger than yours. Every category keyword is contested by well-funded competitors and by the comparison sites that live off them. A smaller vendor cannot outbid them on the head terms; it has to out-target them on the vertical ones.
5. Attribution breaks across the cycle. The click lives in the ad platform, the lead in the marketing automation, the opportunity in the CRM, and the closed deal in the finance system. Unless the click identifier travels the whole way, the channel mix is set on guesswork.
6. Qualification is the product of the funnel, not a filter on it. A lead form calibrated for volume fills the sales calendar with the wrong people. One calibrated for quality starves it. The balance is worked continuously, not set once.

Why it is worth doing on the numbers

When the funnel is built against the cycle, software rewards it, because the customer keeps paying. Clinic to Cloud sells cloud practice-management software to Australian specialists and GPs, a consultative sale with a high qualification bar. In the first four months of the engagement its blended cost per lead fell from A$227.27 to A$96.15 on a 17% budget increase, while monthly lead volume rose 2.76 times. The programme then ran eleven tracked months at a blended A$146.29 across 1,067 leads, and the retainer renewed every month for eighteen consecutive months. The improvement was structural: the same channels, restructured around who was buying and why.

• 57.7%Reduction in blended cost per lead in the first four months, A$227.27 to A$96.15, on a 17% budget increase (Clinic to Cloud)Case study →
• US$14.66Lowest cost per lead across four regulated verticals for an enterprise software vendor, 341 legal-vertical leads at US$0.83 a click (Nuance)Case study →
• A$1,278Cost per signed Microsoft channel partner across a four-year, nine-market recruitment programme, 181 partners on A$231,263 of media (rhipe)Case study →

What follows is the order we now build it: the offer and the website first, then organic and AI search, then paid search segmented by vertical and persona, the audiences that find the committee, the channel-recruitment model for products sold through partners, the conversion flow, the nurturing that carries a lead through the cycle, and the attribution and economics that decide what to do next. Chapter one starts with the thing most software companies skip: what the visitor is actually being offered.

Three offers, three different leads

For Nuance, whose portfolio spans PDF software and speech recognition for legal, medical, government, law enforcement and enterprise buyers, the lead-magnet pathway was designed per product and per persona: free trials where the buyer could self-serve, demos where the sale needed a conversation, whitepapers where the buyer was still framing the problem. The pathway was chosen to match the qualification depth the sales team needed, not the conversion rate the media team wanted. A trial produces the most leads and the least qualified; a demo request produces the fewest and the most; a document sits between and feeds the nurture in chapter seven.

• Trial or freemium. Right when the product can prove itself unattended and the buyer can decide alone. Wrong when adoption needs configuration, data migration or a procurement process; the trial expires before the decision is made.
• Demo or consultation. Right when the sale is consultative and the buyer needs to see their own workflow in the product. It qualifies hardest at the form, so the landing page has to earn the request with proof first.
• Document, benchmark or tool. Right for a buyer who is early, and for the nurture sequence that carries them. Wrong as the only offer, because it never asks for a conversation.

What a software website has to do for a committee

A committee buyer arrives with two competitor tabs open and a shortlist to justify. The pages that matter are the ones that answer the questions the committee will ask each other: what does it cost, what does it connect to, who else like us uses it, and what happens to our data. Pricing that says “contact us” loses the evaluator who cannot get a meeting approved without a number. An integrations page that lists the systems the buyer already runs does more for conversion than a feature tour. Security and compliance pages exist for the person on the committee whose job is to say no.

• Vertical pages. One page per vertical the product sells into, in that vertical’s vocabulary, with that vertical’s proof. Chapter three lands each vertical’s ads on its own page.
• Persona paths. The economic buyer, the technical evaluator and the end user each need a route: the business case, the architecture and the workflow.
• Proof that names names. Case studies with dated, specific numbers, reviews on the platforms the AI engines quote, and logos the committee recognises. Unverifiable claims cost more than they earn.
• The next step on every page. The offer from the section above, in the first screen, with the qualification questions chapter six describes.

Speed and the mobile evaluator

Software buyers research on desktop and check on phones, often from an email on the way to the meeting where the shortlist is agreed. A product page that takes four seconds to render on a phone over a mobile connection is a page that does not get read on the way to that meeting. Page speed is also a ranking factor in chapter two, so it is fixed once and pays twice: measure the pages that carry paid traffic first, on a real phone, and treat anything above two and a half seconds to the largest content as a job before the next flight.

Writing inside the rules

Software sells into regulated buyers even when it is not itself regulated. A healthtech vendor’s pages are read by clinicians bound by patient-privacy law; a product that touches personal data is subject to the Australian Privacy Act, New Zealand’s Privacy Act 2020, the UK GDPR and, for United States buyers, a patchwork of state laws. The practical consequences for acquisition are two: consent for tracking has to be real, which affects what chapter eight can measure, and claims about security, uptime and compliance have to be ones the contract will stand behind. Write the ad the way the sales engineer would say it, not the way the press release did.

Where the evaluator starts

The research path for a consultative software purchase runs from problem to category to shortlist to vendor. Early queries name the problem (“practice management software for specialists”, “speech recognition for law firms”); middle queries name the category with a qualifier (“best”, “for”, “vs”, “pricing”, “alternatives”); late queries name the vendors. The head of the category is held by the comparison and review sites, and the AI engines quote those sites when asked “which is the best”. A vendor rarely wins the head term. It wins the vertical and the comparison terms, and it wins the right to be named when the engine is asked.

• Problem pages. Long, honest pages on the problem the product solves for each vertical, written by someone who has solved it. These are the pages the AI engines retrieve.
• Vertical solution pages. “Product for specialists”, “product for law firms”: the pages that chapter three’s vertical ads land on, and the pages that rank for the “for” queries.
• Comparison and alternatives pages. Fair, specific, dated comparisons against the competitors the evaluator is already comparing you with. Written to be quoted, not to win an argument.
• Integrations pages. One per system the buyer already runs. Low volume, very high intent, and a query the incumbent vendor rarely ranks for.

Pages that earn the click

For each page, the title takes the query and the first screen answers it. A vertical page opens with the outcome for that vertical, in its language, with the proof beside it; a comparison page opens with the honest summary; a pricing page opens with the price. The technical foundation is the same as for any site that wants to rank: one canonical URL per page, a sitemap with real modification dates, internal links from the pages that already have authority to the pages that need it, and page speed measured on a phone. Most software sites fail on the internal linking: the vertical and comparison pages are published and then linked from nowhere.

Being the answer, not just a result

When a buyer asks ChatGPT, Gemini, Claude or Perplexity which practice-management platform suits a specialist clinic, the engine assembles an answer from sources it trusts: review platforms, comparison sites, analyst lists, and the vendor’s own pages when they are specific enough to be cited. The discipline that earns those citations is answer engine optimisation and generative engine optimisation, and it comes down to four things: an entity the engines can verify (a company page that says who you are, where you are, who runs it and what it knows about, with the same facts everywhere the engines look); reviews on the platforms the engines quote; structured data that names the product, its category, its pricing and its integrations; and pages that answer the question a buyer actually asks, in a form that can be quoted. Involve Digital’s answer engine optimisation and generative engine optimisation pages set out the method; for a software company the priority is the review platforms and the comparison pages, because that is where the engines already look.

The same product, four verticals, a 5.6× spread

Nuance’s Dragon speech-recognition products sell to legal, finance, government and law-enforcement buyers, each with a different decision chain and a different tolerance for a cost per lead. Built as segmented funnels per vertical, with vertical-specific messaging, landing pages and lead magnets, the same product returned a cost per lead of US$14.66 in the legal vertical, US$26.17 in law enforcement, US$57.53 in government and US$82.14 in finance. Blended into one campaign those four numbers would have averaged into a figure that justified neither scaling legal nor questioning finance. Separated, they set a budget per vertical, a bid ceiling per vertical and a message per vertical.

• 5.6×Spread in cost per lead between the legal and finance verticals of the same product family (Nuance)Case study →
• 22.63%Conversion rate on PowerPDF search in Australia: one in every 4.5 clicks converted (Nuance)Case study →
• 2.27×Google Search clicks delivered against the contracted volume in the kickoff month, 1,894 against 833, at 47.7% below the contracted cost per click (Clinic to Cloud)Case study →

Structure of the account

• One campaign per vertical, or per vertical and market. Budget, bids and reporting follow the vertical. When markets differ in currency or language, the market splits too: Nuance ran separate media plans for ANZ, EMEA, North America, South East Asia and India.
• Brand separated from generic, product from category. Branded search converts at a rate nothing else will match and tells you nothing about demand generation if it is mixed in. Clinic to Cloud ran branded and specialist-product search as the high-intent capture layer, kept apart from prospecting.
• Phrase and exact match only, with a real negative list. Broad match in a software account buys job seekers, students, tutorials and the free tier of every competitor. Review the search-terms report weekly for the first six weeks, then monthly.
• Competitor terms, deliberately. A small, capped campaign on competitor names and “alternatives” queries, landing on the comparison pages from chapter two. Measured on its own; it will never match generic conversion rates and does not need to.
• One landing page per ad group. The vertical page for the vertical, the comparison page for the competitor term, the integration page for the integration query. Never the homepage.

Bidding for a long cycle

Automated bidding learns from conversions, and a software account with a demo form produces few of them. Two disciplines keep it honest. First, manual or enhanced bidding with a ceiling per vertical until an ad group has enough conversions for the algorithm to learn from, typically thirty in a month, and target-cost bidding after that. Second, the conversion the algorithm learns from has to be the qualified one: a demo request that sales accepted, imported back into the platform, rather than every form submit. Chapter eight covers the import. Without it, the algorithm optimises for whoever fills in forms fastest, which is rarely the buyer.

The contracted numbers tell the story of what structure does to price. Clinic to Cloud’s kickoff insertion order committed to 833 Google Search clicks at A$6.00 a click. The restructured account delivered 1,894 clicks at A$3.14: more than twice the volume at roughly half the price, in the first month, before any conversion-rate work had begun.

Layered audiences, channel-specific roles

Clinic to Cloud’s programme gave each channel a job rather than a share of the budget. Google carried high-intent capture and re-engagement: branded and product search, layered with in-market audiences for business services and computers and peripherals, and customer-match remarketing seeded from existing prospect data, including healthcare executive lists and specialist databases. Facebook ran two motions in parallel: prospecting at scale through lead-generation ads, and mid-funnel remarketing. Display through the Google Display Network carried nurture, against a lead-nurture audience uploaded as a custom-match list. The data work was the connective layer: audience uploads, exclusion lists so that customers and dead leads were never re-bought, cohort analysis of which audiences produced marketing-qualified leads, and a monthly review that fed the next month’s targeting.

LinkedIn for the technical decision-maker

When the persona is defined by job title and the company by size and industry, LinkedIn is the only platform that targets both accurately. It is also the most expensive click in the plan, so it carries the message that needs that precision and nothing else. In rhipe’s Microsoft channel-recruitment programme, LinkedIn carried the Azure-specific targeting at IT decision-makers inside managed service providers, while the volume prospecting ran on cheaper platforms. Use it for the account list and the title list; do not use it for the whitepaper that anyone would download.

Lead-generation ads for market entry

Where a product is entering a market without a website presence, a review footprint or a search history, native lead forms on social platforms produce leads before any of that exists. rhipe used Facebook lead-generation campaigns to carry its South East Asian market entry, producing 1,063 partner leads across eight markets in a single calendar year: 314 in the Philippines, 219 in Indonesia, 127 in Thailand, 105 in Sri Lanka, 88 in Malaysia, 50 in Cambodia, 159 in Australia and one in Bangladesh. The leads were cheaper and less qualified than search leads, which is the point: they fed the nurture and the events in chapter seven while the search footprint was built.

• 1,063Microsoft partner leads from Facebook lead-generation campaigns across eight markets in one calendar year (rhipe)Case study →
• 2.76×Lift in monthly lead volume by month four on a 17% budget increase, 66 to 182 leads (Clinic to Cloud)Case study →
• A$67.50Facebook cost-per-lead floor in the peak month, against a blended A$94.09 (Clinic to Cloud)Case study →

Who not to reach

Exclusions are the cheapest optimisation in a software account. Current customers, lost deals, employees, competitors’ staff, students and job seekers all click on software advertising, and every one of them costs the same as a buyer. Upload the customer list and the closed-lost list as exclusion audiences to every platform, refresh them monthly from the CRM, and add the negative keywords that the search-terms report keeps surfacing. Clinic to Cloud’s exclusion lists were rebuilt as part of the monthly review, not set once at launch.

A high-ticket recruitment play, sustained for years

rhipe is Microsoft’s specialist cloud-licensing distributor in Asia Pacific. Its commercial model depends on recruiting IT services businesses, managed service providers and resellers as Cloud Solution Provider partners, who then build recurring Microsoft cloud revenue through rhipe. Each signed partner is a high-ticket acquisition that goes on to transact for years, which sets a very different cost-per-acquisition bar to a one-off lead. The programme had to deliver a defensible cost per partner across multiple product lines and nine markets, and keep doing so through four years of Microsoft’s programme expansion. Over those four years the Google Ads programme produced 181 click-attributed partner recruitments on A$231,263 of managed media: A$1,278 per signed partner.

• 181Microsoft Cloud Solution Provider partners recruited across the four-year programme (rhipe)Case study →
• A$646Cost per partner in the rhipe Thailand sub-account, roughly half the programme-wide rate, on A$2,585 of media (rhipe)Case study →
• 45%Increase in net new contracted partners across a focused fiscal-year recruitment drive within the programme (rhipe)Case study →

An account architecture that stays readable for four years

The structure was kept deliberately clean: eight dedicated Google Ads sub-accounts, segmented by product, vendor activation and geography. The master account carried the always-on programme in Australia, New Zealand, Singapore, Malaysia, Indonesia and the Philippines; separate accounts carried the content-gated Solutions activity, the Thailand market in its own currency, the USD-billed South East Asian expansion, and the vendor-funded activations for IBM SoftLayer, Red Hat and Acronis. Brand was separated from product, search from display, remarketing from prospecting. It is the structure that makes a four-year programme readable, not the structure that makes a quarterly campaign look tidy, and it is the same principle as chapter three applied at the scale of a distributor.

• Budget rhythm follows the funding. Vendor co-operative funding runs in flights (rhipe’s CSP co-op, Modern Workplace and Azure Go Big programmes); always-on brand and remarketing were funded directly. The account was built so that a flight could be switched on inside it without rebuilding anything.
• Every channel pulling in one direction. Display and video always-on for reach and view-through, retargeting layered on top, social lead-generation for market entry, LinkedIn for the Azure decision-maker, and event registration campaigns timed to the partner summit and roadshows.
• The conversion is the signed partner. Click-through rate and click-to-lead rate both read low in isolation, because the audience is small and the definition is heavy. The number that matters is the cost to bring on a distribution partner who transacts for years.

Events as a conversion surface

Partner programmes convert in rooms. The rhipe South East Asia Cloud Summit registration campaign delivered 109 registered attendees at A$18.36 each, a 5.11% conversion rate from 2,131 Facebook clicks on A$2,001 of media. The registration is cheap because the event does the qualifying: a managed service provider who flies to a summit is a partner prospect the sales team can close. Build the registration campaign against the same audiences as the recruitment campaign, and count the signed partners that follow the event, not the registrations. The programme’s Best Marketing award at the Microsoft Asia Partner Conference came from the same discipline.

The form is a continuous-optimisation surface

Clinic to Cloud’s prospecting ran through Facebook lead-generation forms, and the form itself was treated as a surface to optimise: the field count was calibrated to balance volume against lead quality, a tension that surfaces in every lead-generation programme and is rarely actually worked. Fewer fields produce more leads and more noise for the sales team; more fields produce fewer, better leads and a higher cost per lead that the economics may or may not support. The right setting is found by measuring what sales does with the leads, not by measuring the form, which is why chapter eight’s marketing-qualified-lead cohort analysis fed the form changes each month.

• Ask the questions that qualify. Company size, role, the system they use today, the timeframe. Two well-chosen questions filter more than five generic ones.
• Use the answers. Route a qualified request to a calendar, an early-stage one to the nurture sequence. A form whose answers go nowhere is a form that should have been shorter.
• Native forms for reach, landing pages for depth. The native form on the social platform converts more; the landing page qualifies more. Run both and read the difference in what closes.

Demo, trial or lead form? Test it

The offer from chapter one is a hypothesis until it has been tested against the alternative. For a product with a consultative sale, run the demo request against a shorter “see it for your practice” consultation and against a document offer, on the same traffic, and judge them on qualified conversations a month later rather than on submissions a day later. Nuance’s pathways differed by product and persona because the testing said so, not because a rule said trials suit software.

Speed to lead

A demo request that is answered inside an hour is a conversation; one answered in three days is a competitor’s customer. The follow-up has to be built before the campaign launches: the notification to the right salesperson, the calendar link in the confirmation, the sequence that starts if the call is not booked. In a consultative sale the buyer has usually contacted two vendors; the one who responds first sets the agenda for the comparison.

The gap between the lead and the decision

Clinic to Cloud’s buyer is a practice principal or practice manager evaluating a multi-year software replacement, often with capital approval routed through a healthcare group. That is months, not days, and a lead captured in month one may close in month four. The programme carried leads through that gap with a display nurture layer on the Google Display Network, built on a lead-nurture audience uploaded as a custom-match list, and with Facebook remarketing whose message changed as the lead moved. Neither layer generated leads; both raised the rate at which leads became customers, which is the only reason to run them.

• Separate remarketing from prospecting. rhipe kept them in separate campaigns for four years so that each could be read on its own. A remarketing campaign folded into prospecting flatters the prospecting and hides the nurture.
• Message by stage. Proof for the evaluator, the business case for the economic buyer, the migration story for the person who will run the project. The same advertisement to all three is wasted on two of them.
• Email sequences that respect the cycle. A weekly note with something useful, for as long as the cycle runs, from a named person. Not a drip of feature announcements.

Capacity is part of the funnel

By month four Clinic to Cloud’s programme was delivering 182 leads in a month, and later 186, against a kickoff month of 66. A sales team sized for 66 cannot work 186, and leads that are not worked inside a day become the unqualified leads that make the media look worse than it is. Before a flight scales, ask the sales lead how many qualified conversations a week the team can actually hold, and plan the handover, the routing and the sequence for the overflow. If the answer is that the team is already full, the next dollar belongs in conversion rate and nurture, not in more leads.

• 186Leads in the peak month at a blended A$94.09 cost per lead (Clinic to Cloud)Case study →
• 18Consecutive months of monthly retainer renewal, each one a decision to keep going (Clinic to Cloud)Case study →
• 109Partner summit registrations at A$18.36 each, 5.11% conversion from 2,131 clicks (rhipe)Case study →

Events, webinars and the room

For a consultative sale the event is often where the nurture ends and the deal begins. Registration campaigns run against the same audiences as the lead campaigns, and they are cheap because the event does the qualifying. Count what follows the event: rhipe’s summit and roadshow registrations were judged by the partners who signed afterwards, not by the seats filled. A webinar works the same way at a smaller scale, and its recording becomes the document offer for the next quarter’s nurture.

Rebuild the attribution before the first flight

The click identifier that the ad platform issues has to travel with the lead into the marketing automation, onto the opportunity in the CRM and through to the closed deal, or the channel mix is being set on guesswork. Nuance’s programme was built so that the channel mix could be reset against what actually converted; the practical requirement is a first-party record of the click source, the campaign, the landing page and the time, captured on the first visit and carried through every form the visitor later completes. Consent has to be real for the record to be lawful, and the record has to survive a visitor who leaves and returns a week later.

• Capture on the first visit, not at the form. A visitor who reads a comparison page, leaves, and returns by typing the URL has lost their click parameters by the time they fill in the form unless they were stored on the first visit.
• Carry every identifier. The main search platforms issue more than one kind of click identifier depending on the device and the consent state. Treat the set, not one of them, as the key.
• Write it to the CRM. The opportunity carries the source and the click. When the deal closes, the close can be sent back to the platform against the click that started it.

Define the conversion by the sales stage

The platforms will optimise to whatever is defined as a conversion. Define it as the form submit and they will find people who submit forms. Define it as the sales-accepted lead, imported back from the CRM with its click, and they will find people who become sales-accepted leads. For Clinic to Cloud the marketing-qualified lead, analysed by cohort and audience, was the number that reset the targeting each month; for rhipe the conversion was the signed partner, which is why click-through rates were allowed to read low. Choose the deepest stage that produces enough events for the platform to learn from, and import it.

Close the loop offline

Offline conversion import is the mechanism that carries the CRM stage back to the platform: a scheduled upload of the click identifier, the click time, the stage reached and, where the contract value is known, the value. It is the difference between a platform that optimises to forms and one that optimises to pipeline. It also makes the search-terms report honest, because a term that produces submissions and no opportunities shows up as what it is.

The five numbers to report monthly

1. Cost per lead by vertical and by channel, never blended: the Nuance spread from US$14.66 to US$82.14 is the reason.
2. Lead-to-qualified rate by source: the number that tells you whether cheap leads are cheap or merely unqualified.
3. Cost per sales-accepted lead, and cost per opportunity: the numbers the budget is actually set against.
4. Pipeline created and closed, attributed to the click: closed-won value against media spend, by vertical, for the cohort that started ninety days ago.
5. Capacity used: qualified conversations held against the team’s capacity, so the plan knows when the next dollar belongs in conversion rather than volume.

Set the ceiling from the margin, not the market

The ceiling on acquisition cost is the annual contract value, times the gross margin, times the years a customer is retained, divided by the payback you require. A practice-management subscription held for several years supports a cost per customer that a monthly plan cancelled after a quarter never will, which is why the same cost per lead can be excellent for one product and ruinous for another. Set the ceiling per plan tier and per vertical before the campaign, in writing, and let the media plan work inside it. Then reconcile the ceiling every quarter against the retention and the margin the accounts actually show.

• Customer acquisition cost. All media, all agency and tooling cost, and the sales time to close, divided by the customers closed in the period. Not media alone.
• Payback period. The months of gross margin needed to recover the acquisition cost. Investors and boards read this before anything else.
• Lifetime value to acquisition cost. The ratio the ceiling comes from. It depends on retention, so it is as good as the retention data behind it.

Worked models from the case studies

The pattern across the three is the same. The media programme produces a cost per event at a defined stage; the vendor’s own numbers for close rate, contract value, margin and retention turn that into a cost per customer and a return. The guide can show you the first half with audited figures. Only your accounts can supply the second, which is why the checklist at the end starts with them.

Sensitivity, honestly

A model is only as useful as the range it admits. A cost per lead moves with the season, the vertical and the offer: Clinic to Cloud’s ran from A$227.27 in the kickoff month to A$94.09 at the peak, and the blended eleven-month figure of A$146.29 is the number to plan on, not the best month. A lead-to-customer rate moves with the sales team’s capacity and the quality of the form. State the bands, update them from the closed deals every quarter, and let the ceiling move with them. A return quoted without its assumptions is a number nobody should spend against.

• A$146.29Blended cost per lead across eleven tracked months and 1,067 leads: the planning number, not the best month (Clinic to Cloud)Case study →
• A$94.09Blended cost per lead in the peak month: the number a plan should not assume (Clinic to Cloud)Case study →
• A$231,263Managed media across four years for 181 signed partners: the programme cost the partner economics were judged against (rhipe)Case study →

Growth you can audit

The test of a media programme is whether the customer keeps paying for it. Clinic to Cloud renewed a monthly retainer without interruption for eighteen consecutive months, each renewal a decision made on that month’s qualified leads. rhipe kept its recruitment engine running for four years through the full ramp of Microsoft’s programme expansion, and a focused fiscal-year drive inside it lifted net new contracted partners by 45%. Nuance, one of Involve Digital’s foundation clients, ran hundreds of campaigns across a multi-year engagement. None of those is a promise about your product. Each is the kind of evidence to ask any agency for before you spend.

What to do, in order

1. Write the economics down. Contract value, margin, retention, payback, per tier and per vertical. The ceiling comes from here.
2. Fix the offer and the pages. The lightest offer that still qualifies, one vertical page per vertical, pricing and integrations the committee can find, on a site that loads on a phone.
3. Rebuild attribution. Click capture on the first visit, carried into the CRM, with the offline import ready to send the qualified stage back.
4. Start organic and the entity. Problem pages, vertical pages, comparison pages, review platforms and structured data. Slow, compounding, and the cheapest pipeline you will own.
5. Build paid search by vertical and persona. Phrase and exact, brand apart from generic, one landing page per ad group, ceilings per vertical, bidding to the qualified conversion.
6. Layer the audiences. Each channel with a job; exclusions refreshed monthly; LinkedIn only where the title matters; native forms where the market is new.
7. Tune the form and the follow-up. Qualifying questions, routing by answer, speed to lead, and a monthly reading of what sales did with the leads.
8. Nurture through the cycle and plan the capacity. Remarketing apart from prospecting, message by stage, and a sales team sized for the volume before the volume arrives.
9. Report the five numbers monthly, and reset the mix. Cost per lead by vertical, lead-to-qualified, cost per opportunity, attributed pipeline, capacity used.