HAVAG fare experiment proves profitability of customer reward programmes
Bonus schemes, reward programmes, loyalty discounts, customer engagement and other incentives are designed to motivate occasional passengers to use public transport on a more regular basis. Even though the use of these various schemes is widespread, there is little reliable data on their actual impact. Therefore, when Hallesche Verkehrs AG (HAVAG), a local public transport operator in the German state of Saxony-Anhalt, was thinking about introducing a customer reward programme across all its services, it decided to road-test a range of models first before proceeding with a service-wide roll-out. It limited the trials to its transport services in the city of Halle (Saale) and set itself two objectives: to understand how the discount scheme affects customer behaviour and to ascertain whether the scheme generates enough revenue to offset its costs. The fare experiment ran from 1 April to 30 June 2023. We spoke to Kathrin Jähnert-Elster, Fares and Sales Coordinator at HAVAG, who shared her first-hand experience of this process as well as its initial results.
FAIRTIQ: Why did HAVAG decide to go ahead with this fare experiment even though it coincided with the introduction of the discounted nationwide Deutschlandticket?
Kathrin Jähnert-Elster: The Deutschlandticket is primarily aimed at more regular users for whom spending 49 euros a month on public transport represents good value for money. But there are lots of people who only use public transport occasionally. This is the group we want to reach. We wanted to precisely understand if and to what extent loyalty discounts and reward programmes encourage this group to step up their public transport use.
We want to use the knowledge the experiment generated to design and develop an 'eFare'. As part of the StadtLand+ pilot project, we planned to introduce an 'eFare' in Halle (Saale) between August 2022 and December 2024 before rolling it out across the entire MDV [Mitteldeutsche Verkehrsverbund: association of public transport providers in Central Germany] network. We also wanted to use that time to try out different promotional offers and discount schemes. We used the FTQ Lab (Blue) app for these field trials but continued to offer our standard fare models in the main FAIRTIQ (Red) app.
HAVAG and FAIRTIQ jointly agreed on the test design. Two groups simultaneously tested a range of loyalty discount programmes. Discount thresholds, rates and timings differed for each group:
|
Test group |
Discount threshold (journeys per month) |
Discount rate |
Discount outcomes and timings |
Discount outcomes from passenger POV |
|
1 |
4 |
10% |
... on the subsequent 5th, 6th and 7th journey in the same calendar month |
Frequent use of public transport was 'immediately rewarded', with a reset at the start of the subsequent month, i.e. customers pay the full fare again until they reach the discount threshold. |
|
8 |
15% |
... on the subsequent 9th, 10th and 11th journey in the same calendar month |
||
|
12 |
25% |
... on the 13th and all subsequent journeys in the same calendar month |
||
|
2 |
5 |
5% |
... on all journeys in the subsequent calendar month |
The 'reward' was delayed, but regular use led to constant discounted fares. |
|
9 |
10% |
|||
|
13 |
15% |
FAIRTIQ: Which incentive model do you think is the most promising – direct reward versus discounted fares the following month?
Kathrin Jähnert-Elster: From the customer's point of view, it depends on their individual travel behaviour. We didn't just want to rely on our gut feeling. We wanted to make an informed decision, which is why we tested both.
Around 1,760 customers took part in the field test. We randomly split the test users into two groups and set up a control group as well, who did not receive any discounts. Travel and spending behaviour was the same for all groups at the start of the experiment.
The test users received an email at the start of the experiment informing them of the time-limited reward. Further communication took place directly via automatically generated and customised messages in the FTQ Lab app (nudging). For example, test users were notified shortly before reaching the discount threshold and told of even higher fare discounts if they further stepped up their bus and train use.
FAIRTIQ: What points do you think are particularly worth mentioning in relation to your experience with the FTQ Lab app?
Kathrin Jähnert-Elster: The change in travel/use behaviour that the discounts induced.
The results of the 'immediate reward' scheme tested by Group 1 are unambiguous. Public transport use increased significantly, and participants spent around 20% more on public transport than the control group. The result is also statistically significant (p < 0.5), which clearly demonstrates that this sliding-scale discount scheme was a success. It's also worthwhile from the user's perspective. Additional passenger spending was roughly twice as high as the scheme's costs.
The results for the 'downstream reward' tested by Group 2 were generally positive but were not entirely statistically unambiguous. A second test with a larger user group might allow us to analyse the impact and effectiveness of this reward model in much greater detail.
FAIRTIQ: The results show that reward schemes can also be profitable. Given these results, what is the next step for HAVAG and MDV?
Kathrin Jähnert-Elster: The test has only ended recently, so our detailed evaluation work is still ongoing. Once we have the results, we intend to present them to and discuss them with MDV. The findings will inform our subsequent plans.
HAVAG's willingness to experiment not only facilitates MDV decision-making but is also a major step towards greater innovation in the public transport industry. Sound measurement methods produce reliable results and generate valuable knowledge and insights for the entire sector. Because the FTQ Lab app delivers hard data rather than hunches, it opens up the possibility of developing a fare system at low cost and testing what solutions are best suited to the given context and objectives.
You can read more about the HAVAG fare experiment and its findings in our case study.