I have over a decade in B2B/B2C marketing, specialising in CRM, Lifecycle Strategy, and Multi-Channel Campaign Management β email, SMS, WhatsApp, webinars, and beyond β with a consistent focus on lead nurture, retention, engagement, revenue driving & re-engagement.
I untangle operational knots, build the systems & workflows that others rely on and still care about what lands in the inbox. I’ve consolidated complex systems down into manageable tools, delivered huge email suite overhauls a full quarter early, and built frameworks & ways-of-working using AI that have fundamentally changed the pace at which I, and my company, can work.
I pride myself on being highly organised, endlessly curious, having an eye for clean design, and being tuned into the logic-leap between good data and great decisions. A sentence on me: exercise mad, 20+ years in a brass band, a nearly 10-year Duolingo streak, a passion for craft ale, thousands of photos posted on my personal website, and a 2:57 Manchester Marathon.
CRM & Email Marketing Manager
Higher Ed Partners
Email engagement
Enrolment funnel
- Own CRM & lifecycle strategy for 12+ partners across multiple countries and languages β 400k+ prospects, 70+ concurrent programmes, and 1,400+ emails.
- Led the Salesforce replatform for 12 partners (including multi-lingual), producing company-wide technical documentation used from call-centre to director level.
- Consolidated 220+ engagement programmes to 20 β cutting complexity and cost without losing functionality, with documentation adopted across the business.
- Built hyper-personalised journeys at scale β tailoring content, timing, and messaging by partner, programme, country, and lifecycle stage, which is what drove the engagement and funnel gains.
- Unlocked WhatsApp as a CRM channel β taking ownership of template strategy, compliance, and comms framework from scratch, integrating it alongside email and SMS across full lifecycle programmes.
- Overhauled 400+ emails and restyled 1,000+ templates across 12 partners β refreshing lead nurture, RFI programmes, and agent-facing layouts; lead nurture delivered a full quarter ahead of schedule.
- Built and launched survey and testimonial programmes β 700+ student responses and 200+ consented testimonials, surfaced in a bespoke auto-updating analysis dashboard and used directly to shape product and marketing strategy.
- Managed, trained, and developed a LATAM-based CRM Executive β setting clear KPIs and performance tracking that freed up strategic bandwidth and increased execution speed across CRM and campaigns.
- Designed bespoke internal tools from scratch β a live CRM Stage Viewer and Survey Reporting Workbook, built using APIs, Claude Code, and automated pipelines; replacing manual processes and making real-time data accessible across the business.
- Embedded AI into daily CRM practice β custom workflows, prompt frameworks, and functional tools that compress multi-day tasks into single sessions and moved Q4 deliverables into Q1.
Case studies β the problem, what I did, what happened, and what I learned
Lead Nurture 2.0 Rebuilding the emails that turn enquiries into students
About 100,000 people who had enquired were getting our follow-up emails. Over time the setup had been copied again and again, so every partner had its own copy of every list and every email flow. Deadline emails often pulled from fields that were empty, and nothing was joined up with our phone calls or texts. When we went through the journey as if we were a student, some partners sent nothing after the first ‘thanks for your enquiry’ email, and others called five times in two weeks.
I ran it from start to finish, in three stages.
- New emails. Three general emails and two deadline emails for each stage: when we were still trying to reach someone, when they had not started an application, and when they had started but not finished. A deadline email always comes before a general one.
- A map of every message a student gets, built from that walk-through, then a check of every adviser email template, partner by partner, for consistency, the rules and layout.
- A simpler setup underneath. Three shared lists that pull from the most reliable data in the CRM, feeding three email flows that split by partner, instead of a separate flow for every partner.
- A careful launch. One partner first, the others added once each was checked, and the old emails switched off in the same order.
| Before | After | |
|---|---|---|
| Enquiry to successful contact | 67.1% | 74.7% +7.5 pts |
| Enquiry to enrolled | 50 days | 36 days 28% faster |
| Applications started | 43.0% | 43.9% +0.9 pts |
| Applications submitted | 21.2% | 23.0% +1.8 pts |
| Old | New | |
|---|---|---|
| Delivered | 94.9% | 98.8% |
| Hard bounces (address doesn’t work) | 4.30% | 0.57% |
| Opened | 48.7% | 56.5% |
| Clicked | 1.68% | 3.01% |
The biggest gain came where we split one group of people into two, based on what they actually wanted. Clicks at that stage went from 1.21% to 3.85%, and bounces fell from 5.37% to 0.49%. Some old emails bounced for 15–24% of the people they went to. No new email goes above 5.5%, and nearly all are under 1%.
These numbers compare the four months before with the four months after. It was not a controlled test, and other work was happening at the same time, so they show which way things moved, not proof that this alone caused it. Two results went the wrong way, and they taught me the most:
- The first email in the journey gets opened six points less than the old one. Most likely because it now reaches people who have only just enquired, not people who had been on our lists for months. I have not proved that yet.
- Emails written as ‘I’ instead of ‘we’ lost 11.3 points of opens in a test of 3,700 emails, with five of the eight clearly worse. I advised against using them, and redesigned the test to keep the subject line the same, so we can see the effect of the writing style on its own.
220 programmes down to 20 One setup for all twelve partners, changed in one place
The old setup treated every partner as if it were a separate company. Each email journey was built once for every partner, each with its own lists of who gets it, who is left out and who is checked. That came to about 220 email flows, and far more lists. Changing one line of text or a send time meant doing it up to twelve times, with no easy way to check every copy had been done. The deadline emails also relied on data that was often missing, so the most important emails could fail without anyone noticing.
Three things.
- Rebuilt the lists so they pull from the data the CRM actually keeps up to date, based on where each student is in their course. One list now works for every partner.
- One email flow per journey, which splits by partner. The rules, the timing and who is left out are set once; only the partner’s own wording sits on each branch.
- A careful switchover. One partner first, the old flows switched off in stages, and a label on everything new, so old and new can be told apart with a filter instead of by hand.
| Before | After | |
|---|---|---|
| Email flows | about 220, one per partner per journey | about 20, one per journey |
| Payment-deadline journey | 45 parts across nine partners | 3 parts |
| Main follow-up flows | one per partner per stage | 3 |
| Deadline reminders | one per partner per type of deadline | 5 |
| Reporting on one journey | 92 items sorted by hand | one filter |
| Delivered | 94.9% | 98.8% |
| Hard bounces | 4.30% | 0.57% |
| Deadline emails | sometimes failed without anyone knowing | sent reliably, before the deadline |
To be fair about it: the better delivery and bounce numbers come as much from rebuilding the lists on up-to-date data as from cutting the number of flows, and the time saved is written down rather than measured. The real win is the habit. Change the wording once and all twelve partners get it, and adding a new partner means adding a branch, not building everything again.
Making WhatsApp work at scale From 200-message batches to all twelve partners
We had WhatsApp, but could not use it properly. We could only send about 200 messages at a time. Message templates were tied to each partner’s own phone number, so every change had to be made twelve times. Once WhatsApp approves a template you cannot edit it: you have to delete it, build it again, and wait 30 days to reuse the name. We had ended up with 36 live templates, none for the deadline messages students needed, no access to the messaging platform for the marketing team, and no figures to measure anything against.
- Removed the 200-message limit. I led the marketing side of a project to send in bulk from Salesforce, including testing it and checking it worked once live, and set up campaigns for each of the twelve partners with four deadline campaigns each, so sends across partners are planned, not made up on the day.
- Cut the planned 35 templates down to 30, combined the deadline messages we needed, and made the tone and sign-off the same throughout.
- Built 360 templates by hand: 30 for each of the twelve partner accounts. I wrote a small browser tool that fills in each partner’s name and number in the order the WhatsApp form asks for them.
- Agreed a process with IT. No quick one-off edits: changes are requested, reviewed and signed off by marketing. Plus a repeatable way to measure results using WhatsApp’s own stats.
WhatsApp can now reach all twelve partners at scale, with a message ready for every step from enquiry to enrolment.
- The 200-message limit is gone. Bulk sends run through Salesforce, with campaigns set up for each partner, so sending to several partners is planned instead of squeezed into batches of 200.
- Every step has a template. The deadline messages did not exist before. They do now, on all twelve numbers, alongside the enquiry and on-course messages.
- 30 templates live, built 360 times by hand with a tool written for the job.
- A clear way to add new messages. Because the process with IT is agreed, a new message is a simple request, not a rebuild.
Measuring WhatsApp properly for the first time showed three things worth acting on:
- Low clicks were mostly a measuring problem. Every message asked people to type a reply, and WhatsApp does not count a typed reply as a click. So we did not really know how many people responded, and adding tap-to-reply buttons became the most useful next change.
- Messages that did not arrive were down to old phone numbers in the CRM, mostly at three partners. A data problem, not a WhatsApp problem.
- The same five templates did badly at partners with nothing in common, which means the problem was the wording, not the audience. That is what the rewrite went after.
Each of these gives a clear next step with a reason behind it, not a guess. That is the other benefit of running the channel properly.
Student Voice From no student feedback to live data
We had no proper way of hearing from the students we recruited. Nothing told us, across partners, whether a module met expectations or how happy students were when they finished. Testimonials came in at five to eight a quarter, from wherever we could find them. Talks with partners about quality relied on stories, there was no early warning that students might drop out, and reports, when there were any, were slides built by hand from small spreadsheets and already out of date.
Four surveys at different points in a student’s course, and the reporting to make them useful.
- Mid-module survey: sent when students are most likely to drop out, asking if the module met their expectations and what would help.
- End-of-course survey: how happy students are when they finish.
- Testimonial request: sent automatically, asking for permission to publish and a score out of 10, so a quote can be used without going back to ask.
- Non-starter survey: the newest, sent to people who enquired or applied but did not enrol, asking why and what would have changed their mind.
- The reporting: a live workbook anyone can open, which always shows the latest data; a tool that reads the survey exports, groups comments by topic and matches each course to its subject area; and an automatic update so nobody has to do it by hand. We also agreed some ground rules, such as quoting students instead of averaging small numbers.
| Before | Latest | |
|---|---|---|
| Mid-module responses | none | 524, across all twelve partners |
| End-of-course responses | none | 288 |
| Testimonials | a few, 5–8 a quarter | 186: 160 with permission, 103 scored 8–10 |
| Reporting | the odd hand-made slide deck | live workbook, updates itself |
| Opened | Replies per 100 sent | |
|---|---|---|
| Mid-module | 74% | 6.3 |
| End of course | 79% | 12.4 |
| Testimonial | 64% | 2.4 |
Overall, 76% of students rate their module well, and 82% rate their whole course well.
None of this was known before:
- Module two is where it dips: 3.74, against 4.20 and 4.26 for the modules either side. That tells us exactly when students need more support.
- Two subject areas score well below the rest. That is a conversation about the courses, not the marketing.
- Almost every testimonial comes from one email: 72 of 74 last year. Over a hundred students who gave five stars at the end were never asked for one.
That bank of testimonials, with permission to use them, is also what made subject-specific testimonials in Lead Nurture 2.0 possible.
Senior Marketing Executive
Parkingeye
- Owned and continuously improved the CRM, CMS, email marketing & customer journey strategy across the UK and sister brands in Germany, Austria & Denmark.
- Launched a competitor charging brand from zero β handling CRM installation, website, marketing, branding, and global stakeholder liaison, scaling the database from 0 to 500k.
- Drove improvements & efficiencies across 5 brands in 4 countries, delivering consistent YoY growth throughout.
Earlier experience