Manual Focus · Light Meter · Rider Panel

The Rider Panel: a simulated world of 500 AI agents

The Rider Panel tests how buyers respond to a new product before it launches. For this first study we recruited 500 UK cyclists who will shop for an indoor trainer between September and March, from a potentially limitless pool of agents the Rider Panel can draw on. Each one has a life your launch has to fit into, with a family or a flat to think about, kit they already own and a budget to stick to. They live the season month by month. They choose from the trainers someone like them would actually come across, at real UK prices.

One agent can end the season up to 86 different ways, buying one of 12 trainers or other options in any of seven months, or not buying at all. Five agents can end it 4.7 billion ways. For all 500, the number of possible outcomes is about 970 digits long.

500 agents as a knowledge graph, connected by what they have in common

Hover over a dot to see who it is. Click to see their connections.
Pinch or ⌘ + scroll to zoom · drag to move · double-click to zoom in
Use ⌘ + scroll or pinch to zoom the graph

All 500 agents, by name

Pick a riding type, or show everyone. Choose any name to read their life story.
What the study found

The results of four launch plans for a new smart trainer

We invented a brand called Northfold and a folding smart trainer, the DD1. Northfold isn't a real company. The DD1 turns your own bike into an indoor exercise bike. It's quiet and accurate. It folds flat to 14cm deep for small homes. Like its main rivals, it comes with the Zwift Cog and Click. It goes on sale on 16 November at £499. Sigma Sports and Tredz stock it, with a display stand in Halfords.

We ran the same season four times on the same 500 agents. Their lives stayed the same between runs. So did the weather and every rival's price. Only the Northfold launch changed, so any difference in what they bought comes from that one decision.

Ranges come from re-sampling the 500 agents 4,000 times. A change counts when its range stays clear of zero. Revenue is Northfold's revenue from the 500 agents.

Why agents decided

Every agent gives the two or three things that decided it for them, in their own words. For a brand, this turns a sales number into a to-do list.

Why agents bought the Northfold

Why agents walked away

From the £499 launch. An agent can give more than one reason.

Agent diaries

Four agents' diaries through the season

Every agent keeps a diary from the month they start looking. Each month includes the real weather in their town. Compare an agent who bought at £499 with one who waited for Black Friday.

    Who the 500 are

    Who the 500 agents are

    The panel isn't every cyclist. Most cyclists never think about an indoor trainer in a given year. These 500 are the ones who go shopping for one this season. We built 200,000 UK cyclists from survey data first. Then we gave each of them a chance of shopping this season based on what they ride indoors now and what's wrong with it. The 500 were drawn from everyone who shops, so their mix follows from the simulated population. About 2 in every 100 adult cyclists shop for a trainer in a season, on our estimate. Four in ten of these agents are buying their first, which is our assumption. The sources and working are further down this report. Five hundred was our choice for a first study. The pool has no limit, so a brand's study can be as large as its question needs.

    Each agent is a simulated person, not a persona on a slide. Their life story covers work and family. Their profile records the bike computer they own and how much its brand matters to them. It also records the cassette on their bike and where a trainer would live at home. Their kit never appears in their life story, so it can't take over the plot. This first panel is set in the UK. For another country, every input comes from that country instead.

    Life story and household

    Who they are follows UK survey data. Each agent also has a written life story with a reason to start looking this season.

    Sport England Active Lives · English Housing Survey · FCA Financial Lives

    Real weather in their town

    Every agent lives the actual weather their town had in the same months last year. Rain on the commute. Frost on the roads. Dark before most people finish work.

    Open-Meteo weather archive · 51 UK towns · September to March

    Retail calendar, September to March

    The clocks going back. Black Friday from 20 to 30 November. Christmas costs. The January sales. Spring arriving in March.

    UK retail calendar · search seasonality

    10 real rival trainers at UK prices

    Ten real rival trainers at their UK price on the day, with their Black Friday and January deals. Nobody knows a sale price before it starts.

    UK retailer pages · 28 October 2026
    500in-market agents, each deciding for themselves

    Which trainers each agent knows

    Each agent only knows the trainers someone like them would meet. A Garmin owner sees Tacx in Garmin Connect. An agent who shops at Decathlon sees Van Rysel. Brand familiarity follows real UK search share.

    Set in code from what each agent owns and where they shop

    Your launch plan

    Your price and where it's sold. Your reviews. How many agents your launch reaches, capped at what an established challenger brand achieves.

    The only thing that changes between runs
    Meet the agents

    Eight of the 500 agents

    These eight are a sample. Every agent in the panel has the same depth: a written life story, a reason to start looking and a worry this season. Choose a card to read the full story.

    How a study works

    How a Rider Panel study works, in four steps

    1. Draw 500 agents from survey data

      Draw the people who shop for your category this season from a simulated population built on survey data.

    2. Load the real market

      Load each town's real weather with the retail calendar. Add every rival at its real price.

    3. Write each launch plan

      Your product and price. Your shops and reviews. Write each version of the launch you want to compare.

    4. Run the season once per plan

      Every agent lives the season month by month and keeps a diary. The difference between runs is what your decision changes.

    Accuracy checks

    How we checked the results

    We set the pass marks before any results came in, so we couldn't move them afterwards. Here is every check, including the two the panel narrowly missed.

    Sales in real units

    What the results mean in real sales

    Each agent stands for about real UK shoppers. That's set so the £499 run's trainer sales match the size of the UK market this season. Our market figure is an estimate, so put in your own if you have one.

    Our estimate is about 185,000 indoor trainers a year, worked out from market revenue, with low confidence. September to March carries 65% of a year's UK trainer searches, which gives about 120,000. See the working.

    Northfold's first-season reach is capped at the awareness of JetBlack, an established UK challenger brand. A new brand can't be better known than that in its first season. The baseline awareness figure is a setting in the simulation, not a survey result, and the working below shows how it was set.

    Northfold DD1 sales, September to March

    Ranges come from the panel only. They don't include doubt about the market's size.

    Sources and working for the market figures

    We found no published figure for UK trainer sales or ownership, which means several numbers in this report are our own estimates. Each card gives the source or the sum, so you can check it or swap in better data.

    The number of adult cyclists

    About 7.4 million. Sport England counted 7.42 million adults who cycled at least twice in 28 days, in the 12 months to November 2025. That's 15.5% of adults in England. The count covers England only because we found no matching UK figure.

    Sport England Active Lives, Table 6

    Indoor trainers sold in a year

    About 185,000, with low confidence. We found no published UK sales figure, so we worked it out from revenue. Dataintelo puts Europe's trainer market at $0.46bn in 2025. We assumed the UK is 12.5% of that, which is about £44m at $1.30 to the pound. An assumed average price of £225 makes about 196,000 trainers. Taking off 10% for repeat buyers and gifts leaves about 176,000 buyers, or 2.4% of 7.4 million cyclists. We rounded that to 2.5%, which is about 185,000. Our low and high assumptions gave 126,000 to 239,000. We widened that to 110,000 to 260,000 because the inputs are weak. Dataintelo's own growth figures don't quite add up, so treat its revenue as rough.

    Dataintelo market report · UK share, price and repeat buying are our assumptions

    Indoor trainers sold this season

    About 120,000, with a range of 71,000 to 169,000. September to March carried 65% of a year's UK Google searches for five generic trainer terms, measured from October 2025 to September 2026. Applying that to the yearly figure gives 185,000 × 65%, which is about 120,000. Searches stand in for sales because we found no monthly sales data.

    DataForSEO search volume · Five generic trainer terms

    The share of cyclists who shop in a season

    About 2 in every 100. In the £499 run, the 500 agents bought 422 trainers. Dividing 120,000 by 422 means each agent stands for about 284 real shoppers, so the 500 stand for about 142,000. That's 1.9% of 7.4 million. The sales figure is for the UK while the cyclist count is for England, so the true UK share is a little lower. The rules we used to build the panel had 3.0% shopping, or 5,905 of 200,000. We set those chances ourselves, so we quote the figure tied to the market estimate.

    Our working · Panel results with the estimates above

    The share buying their first trainer

    Four in ten, or 200 of the 500. These agents have no trainer or smart bike at home now. This is our assumption, not a measurement. We set each cyclist's chance of shopping so about 40% of shoppers would be first-time buyers, because we found no UK survey of trainer ownership.

    Our assumption · No UK ownership survey found

    Brand familiarity

    Each brand's share of UK Google searches across the five trainer brands in an average month, measured in October 2026. Wahoo has 58%, Tacx 19%, JetBlack 15%, Elite 5% and Van Rysel 3%. Searches show how often people look for a name. They don't measure sales.

    DataForSEO search volume · October 2026

    The awareness of JetBlack

    119 of the 500 agents have JetBlack on their shortlist. That number comes from our panel, not a brand awareness survey. We set the chance of knowing JetBlack at 35% for keen riders and 12% for everyone else, with its search share as a guide. We then scaled Northfold's reach so a strong launch could meet about 120 agents. A brand's own tracking data or media plan can replace this.

    Our assumption · Guided by DataForSEO search share

    Who the simulated cyclists are

    Sex and riding purpose follow Active Lives. Regions in England follow DfT weekly cycling rates multiplied by ONS adult populations. Outdoor space and renting follow the English Housing Survey. Buy now pay later use follows the FCA's Financial Lives survey. Most of these sources cover England only. Scotland and Wales, age, income, riding type and indoor setup are our own assumptions, informed by DfT travel surveys where we could.

    Active Lives · DfT CW0302 · English Housing Survey · FCA Financial Lives
    What to do next

    What Northfold should do next

    Each action comes from what the agents did and said. Three were tested by running the season again. The others are next in line.

    Why it matters

    How the Rider Panel compares with focus groups and surveys

    In our experience, asking real customers is slow and expensive, so most brands only do it for their biggest launches. Everything else gets decided on instinct. The Rider Panel makes testing a launch cheap enough to do every time, before you spend money on stock and campaigns.

    Focus groups and surveysThe Rider Panel
    Time to an answerWeeks to recruit and runA few days
    Testing what-ifsOne version per study. Each change means a new study.Change the price or a channel, then rerun the same season on the same people
    CompetitorsHard to show a realistic market in a roomEvery rival at its real price, including its Black Friday deal
    TimingPeople guess what they'd do in three monthsAgents live the season month by month, in real weather
    Why people decideWhat people say in front of othersEach agent's own reasons, plus what would change their mind
    Real customersYesNo. It tells you which plans are worth taking to them.
    Run it on your product

    Run a Rider Panel study on your next launch

    Rider Panel studies are part of Light Meter. We run them for cycling brands in the UK. We can build a panel for any other country on request. Send us the product with the plans you're choosing between. You get back what each plan does against the real competition. Every agent's reasons come in their own words, along with what would change their mind.