Three more live forms
6 live forms · portal 20114015
| # | Form | What I'd change |
|---|---|---|
| 1 | Marketing form /en/contact-sales/ | The only form that already asks the three things routing needs, so use its field list as the template for the rest. One fix: the charge-point dropdown offers 101-500, but your pricing splits Starter and Advanced at 200, so anyone picking that band could be either tier. Split it into 51-200 and 201-500. |
| 2 | Whitepaper download /en/downloadables/… | Two changes. First, ask for more than an email, spread across visits so registrations don't drop: company and country on someone's second download, business type and charge points on their third. Second, add a hidden field recording which whitepaper was taken. All 11 report as one form today, so you can't see which asset actually brings in leads. |
| 3 | Webinar registration /en/events/… | Same staged questions as the whitepaper. Then replace the 13 separate forms with one, plus a hidden event name field. Every webinar is its own form ID today, so nothing rolls up. |
| 4 | Starter application /en/starter-plan/ | Add one dropdown: do you operate a charging business? Without it a property owner and a charge point operator fill in an identical form, and those two belong in opposite places. One goes to a partner operator, the other to sales. |
| 5 | Product interest /en/products/monta-ai/… | Someone asking to try the product is the warmest lead on the site, and this form asks for less than the newsletter does. Add charge points and business type here before anywhere else, and fire a same-day task to an SDR. Intent this high goes cold in days. |
One field fix, one property audit, and one progressive-capture strategy
_v3 and _v4number_of_cps_v3 and
inbound_form_monta_customer_type_v4. The _v3 and _v4 suffixes suggest
versioned properties. I'd audit their history and dependencies before consolidating anything.
The obvious fix is to put eight fields on the webinar form.
I wouldn't.
Instead I'd use progressive fields. You give the form a ranked list of questions but it only ever shows
three at a time. When someone returns, HubSpot recognises them from the tracking cookie, hides what it already
has answers to, and shows the next three. The form always looks short. You learn more each visit.
| Which visit | What the form asks for |
|---|---|
| 1st — the webinar or whitepaper | Email only. Unchanged, so conversion doesn't drop. |
| 2nd — any other gated asset | First name, last name, company, country |
| 3rd — any other gated asset | Business type and charge points — routable from here |
| Thank-you page, every time | An optional "Book a demo", pre-filled with what's already known |
Contact · created 2 minutes ago from Marketing form (website)
Head of Property
mette@nordiskejendomme.dk
Copenhagen, Denmark
Contact-based · re-enrolment on qualification change
The seven sample leads, run through WF-03
| Lead | Channel | Model | CPs | Bucket | Deciding rule |
|---|
240 charge points looks like a large deal. It isn't. Nordisk owns buildings, it doesn't run charging, so there is nothing to sell them.
A council looking now, with a tender next year. None of the four buckets fit, and disqualifying it throws away a deal you had twelve months' warning about.
So it keeps a real bucket — SDR — and takes a status: Watchlist, with a date to come back.
One charger at home. Not a sales lead — but still a Monta customer, just for the app rather than the platform. Send him to Monta Charge and keep him out of B2B email.
If consumer submissions are a meaningful share of the 280, this redirect removes real manual work. I'd measure the actual share first.
Two webinars, eight emails opened, and the only thing on their record is an email address.
As far as the routing is concerned, they're a stranger.
What happens after a lead lands in a bucket · WF-04, WF-05, WF-10
Goal: lifecycle reaches Sales Accepted. A lead nobody touched becomes a marketing problem again.
Self-serve directory
monta_referred_operator.Enterprise assist
Tier note: the partner referral record is a custom object, which needs Enterprise. On Professional I'd track the referral as a deal on its own pipeline, or as three contact properties — referred operator, referral date, outcome. Less clean, same closed loop.
Leads will go stale. The system needs a defined moment where marketing takes them back, rather than letting them rot in someone's queue.
The lists that make it visible
| SDR — unworked past SLA | Escalation |
| Forward to Operator — no outcome | Quarterly reconciliation |
| Watchlist — due for re-qualification | Becomes an SDR task |
Build-ready workflow specification for portal 20114015 · none of it is installed yet
Contact-based · enrols on any of the 6 forms · re-enrols on every submission
Enrols when business model is unknown AND profiling stage < 3
This is the honest answer to "route all 280 automatically". Some can't be routed yet. The job is to shrink that group on purpose rather than pretend it's empty.
Enrols when business model is known AND bucket is unknown · re-enrols on change of model, band or operates-charging-business
Business model decides the bucket. Size only decides who handles it.
Goal is lifecycle = Sales Accepted.
Self-serve directory
Enterprise assist
The referral record is a custom object, which needs Enterprise. On Professional: a deal on its own pipeline, or three contact properties.
Today this is a directory page and a "contact them" link. Once they leave that page you know nothing.
This is how GreenFleet reaches sales without ever filling in a long form.
Someone has to watch the tender portals. That's a process, not a workflow, and I'd own it.
Leads will go stale. The system needs a defined moment where marketing takes them back rather than letting them rot in a queue.
Two proposed lifecycle stages · nine lists, active unless noted
| Stage | Set by |
|---|---|
| Subscriber Gave us an email, asked for nothing else | Newsletter form |
| Lead Submitted any form with intent behind it | Any gated asset |
| Engaged Lead new Score ≥ 25, or two or more touches | WF-08 |
| Marketing Qualified Lead Routed to SDR, or score ≥ 50 with known business model | WF-03 / WF-08 |
| Sales Accepted Lead new An SDR looked at it and agreed it's worth working | SDR, from the task |
| Sales Qualified Lead Meeting held and qualified | Deal creation |
| Opportunity Deal past Appointment Scheduled | Deal stage sync |
| Customer Closed won | Deal stage sync |
| List | Used for |
|---|---|
| Inbound — awaiting routing | Near zero, or WF-03 is broken |
| Manual triage queue | Worked daily by marketing ops |
| Low-confidence routed | Weekly accuracy review, under 80% |
| SDR — unworked past SLA | Escalation |
| Forward to Operator — no outcome | Quarterly partner reconciliation |
| Starter — pending review | Application queue |
| Watchlist — due for re-qualification | Monthly, becomes an SDR task |
| Consumer suppression | Static. Suppressed from every B2B send. |
| Profiling incomplete | Progressive profiling isn't working |
Two processes · the AI extracts, deterministic rules decide
The one designed in detail. Runs as a Custom code action inside WF-03.
| Trigger | Any form with free text, before the routing branch |
| In | lead_info_long plus the structured fields |
| AI returns | Business model, charge points, sites, timeline, country — each with a verbatim quote, a confidence, and any conflicts. No bucket. |
| Rules write | Bucket, sub-motion, lead status, reason, next action |
| On failure | Returns Manual triage. Never a guess. |
Try it on the workflow slide — click the orange Custom code step.
Fires inside WF-04, after routing.
| Trigger | An active SDR lead enters WF-04 — bucket is SDR and lead status is not Watchlist |
| In | Qualification fields and the last 90 days of touchpoints |
| Does | Writes five lines, each linked to the activity it came from |
| Out | Pasted into the SDR task body |
| Guardrail | Never contacts the prospect. Writes "not known" rather than inventing a fact. |
Neither process talks to a customer. Both write to a record a human reads.
| Confidence | What happens |
|---|---|
| ≥ 80 | Routed automatically, sampled at 10% weekly |
| 60–79 | Routed but flagged, 24-hour review task |
| < 60 | Not routed. Manual triage list. Never guessed. |
| Disqualify | Never fully automatic above a charge-point threshold |
Four numbers I'd report
| Straight-through rate | target ≥ 90% |
| Override rate | every override written back |
| False disqualifications | target zero |
| Cost and time per submission | against ~2.5 min manual |
Privacy: only classification fields leave the portal, never contact details. Approved vendor terms, pinned processing region, model and prompt version logged per decision.
Last 30 days · all sources · figures below the MQL line are assumptions, stated as such
Activities stay on the contact, then roll up to the associated company so three people from Nordisk reading the same whitepaper is one account warming, not three leads. The company score is the distinct-touch count, so a shared webinar or a company visit is counted once however many contacts attended. Sales sees account warmth; marketing sees which touches produced MQLs.
| Touchpoint | Captured by | Points |
|---|---|---|
| Industry events | Badge-scan CSV → custom object | 25 / event |
| Website visits | HubSpot tracking code | 2 / session · 10 for pricing |
| LinkedIn paid | Ads integration + conversions API | 8 lead form · 3 click |
| LinkedIn organic | Not natively capturable | 0 — channel level only |
| Whitepaper downloads | Form submission | 15 |
| Webinar | Registration form / attendance | 10 reg · 20 attend |
| Newsletter | Email engagement | 5 click · opens excluded |
Six areas where HubSpot alone won’t quite get you there, and what I’d put alongside it
Registrations are captured. Actual attendance isn't — so one of your strongest engagement signals, someone spending an hour with you, never reaches the data.
I'd either move to a webinar platform that integrates directly with HubSpot, or gate the replay and use replay views as a second signal.
Vector already runs through GTM, but contact-level identification only covers the US. That works for US traffic, and Monta's is mostly European, so its usefulness is limited here.
For European visitors I'd use company-level identification and treat it as an account intent signal, not an attempt to name the person.
LinkedIn gives you no clean way to tie organic post engagement back to individual contacts.
Third-party tools like Dreamdata try to fill the gap, but bring their own limitations and compliance questions.
Today you can see someone reach the operator directory and click through.
If the channel matters commercially: unique tracked links per operator, a simple partner referral record, an outcome field operators can update, and a regular reconciliation.
Otherwise the channel generates revenue and you can't tell which leads produced it.
A standard Event Touchpoint structure linked to both contacts and companies, and the same CSV format for every import.
Neither first-touch nor last-touch tells the story. First-touch credits a whitepaper from a year ago; last-touch credits yesterday's demo form.
I'd use a W-shaped model. It splits the credit across the three moments that actually move a deal — the first touch, the touch that created the lead, and the touch that created the opportunity — roughly a third each, with the remainder spread over everything in between. First-touch and last-touch stay as supporting views.
I'd also group accounts by the month they first entered the funnel and follow each group over time. With cycles this long that usually tells you more than any weighting does.
This is the real lead_info_long field from your contact-sales form. Nobody fills it in the shape a workflow wants.
Engine: not run ·
| Confidence | What happens |
|---|---|
| ≥ 80 | Routed automatically. Sampled at 10% for weekly review. |
| 60–79 | Routed, but flagged. Task to marketing ops with a 24-hour SLA to confirm or override. |
| < 60 | Not routed. Held in a manual triage list. Never guessed. |
| Any, if disqualifying | Disqualification is never fully automatic above a charge-point threshold. The cost of a wrong "no" is a lost enterprise deal. |
Every override is written back to a property. That's the training set for whether the thresholds are right, and the honest measure of whether this works.