Last updated
You've decided to run a mentoring programme. You know roughly what kinds exist - one-to-one, peer, reverse, the rest. So why does the next decision feel so much harder?
Because knowing the types of mentoring tells you what each one looks like. It doesn't tell you how to actually run one. Should a mentor work with one person or six? Should you match people formally or let relationships form on their own? Should this run for a year, or a single afternoon? Those questions aren't about type at all. They're about model - the structure your mentoring runs inside.
This page is about getting that structure right. If you're still deciding which varieties to offer in the first place, our guide to the different types of mentoring is the better place to start. If you've got that part roughly settled and now need to shape it into something that works, read on.
Type vs Model | The Five Dimensions | Recognised Frameworks | Matching Model to Outcome | Models in Context | Blended Models
Type vs Model - and Why Confusing Them Stalls Programmes
Most people use "type" and "model" interchangeably. That's the first place programmes go wrong.
Pick a type - say, peer mentoring - and you've chosen a flavour. You still haven't decided whether peers meet one-to-one or in circles, whether you match them or let them self-select, whether it runs all year or just through induction. Get to launch having chosen the type but not the model, and the programme drifts: people aren't sure how often to meet, relationships can fizzle, and there's no data to show whether any of it worked.
A model is the framework the type runs inside - the rules of engagement. The same type can run inside completely different models, and the model is usually what decides whether the programme delivers. It's also why two organisations running "peer mentoring" can get wildly different results: same type, different model.
(For the related question of whether there's a difference between mentoring and mentorship - subtly yes, and we cover it separately.)
Back to the top
The Five Dimensions That Define a Mentoring Model
Here's the mistake that follows from the first one: treating "which model?" as a single choice from a menu. It isn't. A model is a set of independent dials, and you set each one deliberately.
Direction. Who guides whom.
Traditional mentoring flows from experienced to less experienced. Reverse mentoring flips it. Reciprocal mentoring runs both ways at once.
Ratio. How many people share a relationship.
One-to-one is the classic. One-to-many covers group and team mentoring. Many-to-many describes circles and constellations, where support comes from a network rather than a single mentor.
Formality. Whether the programme is led or left to form.
Formal mentoring is matched, goal-set and tracked. Informal mentoring happens organically. Most strong programmes sit deliberately between the two - formal enough to be fair and measurable, loose enough to let real relationships grow.
Duration. How long a relationship lasts.
Ongoing mentoring develops over months or years. Flash and situational mentoring deliver value in a single focused session.
Delivery. How people actually meet - in person, online through e-mentoring, or blended.
For hybrid and global organisations, the right platform is what makes delivery and measurement possible at all.
Set those five dials and you've defined your model - long before you've worried about what to call it. Even decisions that feel administrative are model decisions: how a programme is licensed and scaled shapes what's possible just as much as the dials above.
CASE STUDY
NHS Clinical Entrepreneur Programme
The world's largest healthcare entrepreneurial training network swapped manual spreadsheet matching for a site-licence model - priced by network rather than per user. That single structural choice is what made matching at scale viable.
Back to the top
Recognised Frameworks Worth Knowing
You don't need the academic literature to run good mentoring, but if you're interested in more in-depth knowledge, a few frameworks are genuinely useful as shorthand.
Developmental versus sponsorship - The developmental versus sponsorship distinction separates mentoring that grows a person's capability from mentoring that actively advocates for their advancement - useful when designing for diversity and progression.
Kram's mentoring functions - Kram's mentoring functions split a mentor's role into career support (sponsorship, exposure, coaching) and psychosocial support (acceptance, counselling, role modelling) - a reminder that mentoring is rarely just about the job.
Constellation model - The constellation model captures the modern reality that no single mentor meets every need, and that people increasingly draw guidance from several at once.
Treat these as viewpoints, not rules. They help you name what you're already trying to build.
Back to the top
Matching the Model to the Outcome
The commonest failure of all: choosing a model because it's familiar, not because it fits the goal. Start from what you're trying to achieve, then work back to the dials.
Chasing retention or engagement? Lean formal and ongoing, so relationships are matched well and supported long enough to take hold. Need depth on a single, urgent challenge? A short situational model beats a year-long commitment. And when match quality itself is the priority, everything comes back to how well your matching is configured.
CASE STUDY
alta / Royal Aeronautical Society
When match quality is the goal, the matching configuration is the model decision that matters most. alta's women-in-aviation programme tuned a bespoke matching algorithm to reach a 99% good-match rate.
Back to the top
The Same Model, Set Differently by Sector
A model isn't fixed - the dials get set differently depending on where it runs.
In education, peer mentoring dominates, usually formal and time-bound around the start of the academic year. In the workplace, models cluster around onboarding, leadership and culture change, often running several at once. In healthcare, coaching and mentoring models frequently run side by side. In associations and professional bodies, cross-organisational and diversity models turn membership into something members actively take part in.
CASE STUDY
A diversity model built on specific matching criteria, run by a professional body: 76% of mentees who completed their profiles found a suitable mentor, accelerating career progression for midwives of colour.
Same dials, different settings.
Back to the top
Blended Models - Combining Rather Than Choosing
The final mistake is assuming you have to pick one. The most mature programmes don't. They run several models at once and let people move between them.
That's only practical when the underlying tooling can hold more than one model at a time - which is increasingly the point of running mentoring through a platform rather than a spreadsheet.
CASE STUDY
East London NHS Foundation Trust
ELFT runs mentoring, coaching and reverse mentoring as three separate programmes on a single platform - each with its own purpose and participants, all matched, managed and measured in one place.
Back to the top
Choosing Your Model
There's no single best mentoring model - only the right combination of dials for your goals, your people and your culture. The skill is in setting them deliberately, then having the tools to manage, nurture and measure what follows.
That's exactly what mentoring software is built to do. Not sure whether it's right for your organisation? There's a free 3-minute quiz that'll help you decide - and if you already know, you can book a demo and see it in action.
