Making business onboarding self-serve
A redesign I led at SSE so business customers can sign up and see all their accounts in one sitting, without waiting days for a manual check.
Much of this work is confidential, so the maps and flows shown here are simplified recreations. Structure and findings are kept true to the original.
Company
SSE (Scottish and Southern Energy)
Type
Service Design
Role
Lead designer (research to UI)
Team
Me, product manager, engineers, business analyst, operations
Timeline
June 2025 to September 2026
How customers felt, before and after the redesign. Hover a phase to compare the two curves.
Problem
Only around 4 in 10 customers who started registering for the online account portal ever finished. The rest dropped out somewhere between typing their email address and getting access.
Usability testing and customer feedback pointed to three recurring complaints:
Telling us things twice. Customers had to re-enter business and personal details they had already given when they signed their contract.
Waiting to be let in. Most sign-ups went to a manual review by a back-office team. That took 1 to 3 days, and up to 4 weeks if extra information was needed.
Missing accounts. Once in, customers often had to request access to each of their other accounts by hand.
“I don’t think it should take 1-3 days for an energy company to open an account.”
Participant, usability testing (paraphrased)
The timing made it worse. People rarely sign up for fun: they register because a bill looks too high, a payment reminder has arrived, or they need consumption data for a meeting or year-end accounts. A multi-day wait lands at exactly the wrong moment.
For the business, every stalled registration meant more calls, more emails and more manual checks, and competitors already offered instant sign-up.
The challenge
How might we give genuine business customers secure access to all their accounts as soon as they sign up, without a manual wait?
Process
My approach
I brought everyone who touches this journey into the same room, because no single team could see all of it. Engineers knew the system rules, support knew what customers said on the phone, and product knew the business goals.
Gathered the evidence. I pulled together usability testing findings, survey and voice of the customer comments, analytics on page views and drop-off, and what support teams heard on calls.
Aligned on the why. In a workshop canvas we described the idea, why it mattered to customers and the business, the situations people are in when they need it, the value we wanted to deliver, and the capabilities and restrictions we had to work with.
Mapped the current journey. Together we built a detailed as-is map covering 15 stages, from the welcome letter to long-term use. Each stage had layers for customer goals, actions, touchpoints, emotions, quotes, pain points, back-office actions and opportunities.
Blueprinted the logic. With engineers and the business analyst, I mapped every system check and decision behind registration, so we could see exactly where and why customers were pushed to a manual review.
Prioritised and designed. We plotted each opportunity on an impact and effort matrix, checked it against security, data quality and development capacity, then I designed and tested the new flow.
Who we designed for
Our primary persona was a business customer for whom energy is only one part of their job: often an office, finance or facilities manager at a business with several sites. They want to get set up quickly, need clear confirmation at each step, and expect their data to be handled securely.
We also kept TPIs in view: brokers and consultants who manage energy for many businesses and need to act on their clients’ behalf.
Areas of focus
Simplicity. Only ask for what we cannot already find out.
Speed. Remove waits, so customers can act on the problem that brought them here.
Visibility. Always show where people are, what comes next and why we are asking.
Security. Keep accounts safe without making every customer prove themselves manually.
The as-is journey
The full map had 15 stages and 9 layers. For this case study I have grouped it into four phases.
- No reasons to sign up
- Links hard to use from paper
- Details already given at contract stage
- Code expiry not shown
- Around 40% never reply to follow-up emails
- Journey stalls here
- Accounts missing on first log-in
- Calls support for help
As-is journey map, simplified from 15 stages to four phases. Switch to To-be to compare.
Process
Feelings dip steadily and bottom out at the manual check, where customers wait days with no clear timescale and many never return.
What the mapping revealed
The manual review was the moment of truth: it was where customers lost momentum, and many never came back.
We asked customers to prove what we already knew. Contract data existed, but registration did not use it, so customers retyped it and staff then checked it by hand.
The wait broke the journey. Around 40% of customers never replied when the back-office team emailed for more information, so their registration simply stalled.
Email alone could not prove who someone was. Only around 5% of customers used an email domain that matched their business records, and stored contact details were not reliable enough to trust. Automation needed a different signal.
Small communication gaps caused big drop-offs. Emails did not say that verification codes expire, did not link back into the journey, and nothing told customers when their supply had started.
Real businesses are messy. One person might manage three separate businesses; one business might need several people to have access. A single login per business did not fit.
“This process takes too long, especially if I am creating an account to fix an urgent problem.”
Participant, usability testing (paraphrased)
From insights to opportunities
We turned each insight into a “how might we” question, then into a design direction the team could build.
How might we verify genuine customers instantly, without relying on email? Ask a few randomised questions only the real account holder could answer, checked automatically against our records.
How might we stop asking customers for things we already know? Prefill business details from a company search and reuse contract data.
How might we give customers every account they are entitled to? Link all eligible accounts automatically once someone is verified.
How might we keep customers informed at every step? Clear progress, honest timings, code expiry messages and links back into the journey.
How might we stay secure without slowing everyone down? Multi-factor authentication for all, with manual review kept only as a fallback for edge cases.
Designing within real constraints
The restrictions we mapped shaped the solution as much as the opportunities did.
Fraud and social engineering risk. Randomised questions, multi-factor authentication and screening of risky email domains.
Unreliable contact data. Verify against billing and supply data rather than stored email addresses.
Complex account structures. Grant access per verified business, so one person can manage several.
A small development team with competing priorities. Phase the work, starting with the change that removed the most manual effort.
Prioritising the work
The team plotted every opportunity by its impact for customers and the business against the effort to build it.
Six opportunities plotted by impact and effort. Hover a point to see which quadrant it sits in.
Process
Quick wins such as clearer messages could ship alongside the big bet: replacing the manual check with automatic verification, which removed the most effort for customers and staff.
The to-be journey
The target journey keeps the same four phases but removes the wait entirely.
The manual check becomes an automatic one, so confidence rises at the exact point where it used to collapse.
- Clear reasons to join
- One tap into the journey
- Contract data reused
- Fewer fields to fill
- Code expiry explained
- Checked automatically in minutes, not days
- Manual review as backup
- Accounts linked automatically
- No access requests
To-be journey map, the simplified target journey.
Solution
The biggest change is invisible: customers now prove who they are in minutes, inside the journey, instead of waiting days for someone to check them by hand.
What changed
Automatic verification. If someone’s details match our records, they go straight through. If not, they answer a short set of randomised questions about their account. Only people who cannot be verified this way are passed to the back-office team.
Less typing. A company search prefills business details, so customers confirm rather than retype.
Accounts linked automatically. Once verified, customers see every account they are entitled to, with no access requests.
Clearer screens and messages. Each screen explains why we are asking, error messages say how to fix the problem, and emails link straight back into the journey.
Secure by default. Multi-factor authentication protects every account, and risky email domains are screened before registration starts.
How verification works now
Customers whose details match go straight through; everyone else gets a chance to verify themselves before a person ever needs to step in.
How verification works now, simplified. Pick a scenario to follow the path.
The new registration flow, recreated without real data. Select a step to see the screen.
Outcome
Impact
Everything shipped in September 2026, and early signs are strong: customers complete registration faster, support teams are seeing fewer onboarding contacts, and customer feedback has been positive.
The first results are in, and the rest are still being measured. Because the journey map is now connected to live data, these figures will update automatically as they come in.
Time to get access. Before: 1 to 3 days, up to 4 weeks. After launch: under 15 minutes.
Customers verified without a manual check. Before: around 5%. After launch: around 80% of accounts.
Customers completing registration. Before: around 40%. After launch: being measured.
Onboarding support contacts. Before: baseline. After launch: falling, being quantified.
Customer effort score. Before: baseline being set. After launch: being measured.
Making journey maps live
A map drawn in a workshop starts going out of date the day it is finished, so I also helped the business set up Cemantica, the journey management platform it had just adopted.
One shared home. Journey maps now live in one place instead of scattered whiteboards.
Live data on every stage. I helped connect data features such as completion rate, customer effort score and page reviews, pulled from sources including Google Analytics.
Health at a glance. Maps update automatically, so a colleague in operations can see how each journey is performing without asking a designer.
Ideas in context. Opportunities raised by any team are logged against the stage they relate to.
Before and after, as seen in the live map
This simplified view shows how the registration part of the journey changed. Customer data and real figures are left out.
The new journey drops the multi-day wait and the steps that asked customers to repeat themselves, and finishes with every account ready.
This turned the onboarding map from a one-off workshop output into a tool the business uses to track the impact of what we shipped.
- Step 1Start registration
- Step 2Verify email
- Step 3Re-enter personal details
- Step 4Re-enter business details
- Step 5Choose customer type
- Step 6Manual review by a team, waits 1 to 3 days, up to 4 weeks
- Step 7Verify phone number
- Step 8Marketing consent
- Step 9Create password
- Step 10Request missing accounts
Illustrative view of the live journey map. Sentiment is assumed, not measured.
Before and after registration steps, an illustrative view of the live journey map.
What I’d do differently
The most valuable thing I did was put engineers, operations and product around the same map: the fix came from connecting what each team already knew.
Data quality is a design material. The obvious solution, matching on email, failed because of the data underneath it. Designing well meant understanding the data as closely as the screens.
Security and simplicity are not opposites. Asking the right few questions felt easier to customers than a long form, and it was safer than a manual check by email.
Back-office journeys matter too. Mapping what staff did behind the scenes showed where customer pain really started.
What’s next: usability testing the security questions, supporting people who manage several businesses under one login, and tracking completion rate and customer effort score as live data in our journey maps.
