trip generation study of drive-through coffee outlets

Trip Generation Study of Drive Through Coffee Outlets

Brian Schapel, Bitzios Consulting

The need for this study

There has been a dramatic increase in the number

of drive-through coffee outlets in recent years

WHY?

Are we working too hard?

Staying up late at night?

We don’t want to get caught napping on the job!

!

Let me repeat that…...

We don’t want to get caught napping on the job!

The need for this study

The RMS Guide to Traffic Generating Developments (Guide) does not yet include

drive-through coffee outlets

Unique operational characteristics compared to other drive-through facilities:

− Mostly limited to coffee, minimal food sales

− No seating for most outlets and limited parking

− Better and consistent planning outcomes – reliable trip generation and parking

demand data

Study scope

Determine the sample number of outlets required to provide meaningful results

Identify suitable outlet survey sites

Obtain agreements from outlets to conduct surveys

Gather site operational data

Conduct on-site surveys to collect all road traffic trip generation data

Tabulate, analyse and graphically present the collected data to identify key

statistical dependency relationships

Recommend traffic generation rates to adopt in the Guide

Site selection

Wide variations in the location, type and operation of outlets

Outlets were sought in metropolitan, sub-metropolitan and regional areas of New

South Wales, Queensland and Victoria

22 outlets were identified as potentially suitable sites

10 outlets provided agreement for surveys

Challenges in getting agreements

− Relatively small businesses compared to large drive-through fast food outlets

− Many very unwilling to cooperate, concerned with business viability, previous

complaints and/or commercial confidentiality

− Lengthy process, in some cases up to two months

Survey procedure and schedule

Sites were surveyed between 12th May 2015 and 23rd June 2015

2 outlets were surveyed for 6 days

− One of the six-day surveys conducted over 12 hours (6:00AM to 6:00PM)

− The other six-day survey conducted over 4 hours (6:00AM to 10:00AM)

8 outlets were surveyed for 1 day on a Tuesday or Wednesday

Morning survey 6:30AM – 9:00AM (2 ½ hours)

Afternoon survey times varied due to differing PM business opening times (2 hours)

Almost all outlets are closed on Sundays

Data Collection – Site Information

Outlet’s physical structure and operation

Building area

Opening times

Number of employees on a typical shift

Product range

Years of operation

Surrounding land use

Relevant local issues

Data Collection – On-site Surveys

Number of site entry and exit points

Frontage roads’ AM and PM peaks

Drive-through lane capacity (length available for queuing)

On-site parking availability (including for bicycles)

Number of waiting bays

Seating provision - internal and external

Number and type of ordering booths or terminals and collection points

Record of the time that a vehicle enters the site

Record of the time that the same vehicle exits the site

Data Collection – On-site Surveys (Continued)

Number of entering and exiting vehicles (cars/HVs) (15 minute blocks)

Number of vehicle occupants (15 minute blocks)

Number of pedestrians and cyclists (15 min blocks)

Number of queued vehicles (every 5 minutes)

Number of on-site parked vehicles relevant to the site (every 15 minutes)

Significant amount of data collection presented challenges for site surveyors as site

layout restricted visibility in many cases

Data Collection – Passing trade

Selected customers were asked three brief questions:

Was the trip just for coffee or had they had dropped in on the way somewhere else

What they were ordering

Their postcode

These questions were aimed at:

Determining trip origin to assist with determining direction of travel in AM

Percentage of passing trade

Establishing a relationship between order size and service time

Preliminary Analysis

Initial data analysis indicated AM period significantly more trips than PM and

unnecessary to undertake further detailed analysis for the PM period

Comparison of daily totals for six-day surveys showed no clear indicator of which

weekday is the busiest

Saturday is less busy than the week days.

Only three outlets had any internal or external seating, therefore parking analysis

unreliable. Limited available parking and maximum was 8 parked vehicles.

Survey data and key derived statistics were cross-checked for expected

consistencies and variations against:

− RMS Guide to Traffic Generation Developments;

− Land Use Traffic Generation – Data and Analysis 22: Drive-Through

Restaurants (1993)

− Land Use Traffic Generation – Data and Analysis 5: Fast Food (1980), and

− ITE Trip Generation Rates – 8th Edition

Preliminary Analysis (Continued)

Trip rates contained in the RMS Guide for KFC and McDonalds and Institute of

Traffic Engineers (ITE):

Survey RMS ITE

AM Site Peak AM Site Peak AM Network Peak AM Network Peak

DCO’s KFC McD KFC McD Coffee W/- Drive-through

105 150 260 100 180 102

Data Analysis - Methodology

Relationships between variable independent and dependent data tested to

determine statistically relevant linkages between various parameters and the drive-

through trip generation

Initial analysis of survey data showed no significant association between variables

Simple linear regression analysis was conducted to derive R2

R2 represents the percentage of variation in the dependent variable

Values less than 0.80 (80%) not considered accurate enough to indicate a

significant relationship between the dependent and independent variable

Data Analysis - Results

Key relationships tested for R2 to establish key influences on trip generation and queue

lengths (dependent variables) as a priority

R2 results of the linear regression testsIndependent Variable Dependent Variable Reference R2

Frontage Road Network AM Peak Hour Trip Generation Sec. 5.2.1 0.14

Frontage Road Site AM Peak Hour in CBD Direction Trip Generation Table 2 0.12

Frontage Road Site AM Peak Hour Queue Length Table 3 0.26

Frontage Road Two-Way Network AM Peak Hour Trip Generation Sec. 5.2.1 0.12

Gross Floor Area (GFA) Trip Generation Table 4 0.01

Site AM Peak Trip Generation Queue Length Table 5 0.67

Number of Staff Service Time Table 6 0.64

Number of Staff Trip Generation Table 7 0.31

Service Time Queue Length Sec. 5.2 0.07

Service Time Trip Generation Sec. 5.2 0.07

Number of Service Booths Service Time Sec. 5.2 0.06

Number of Service Booths Trip Generation Table 8 0.61

CBD In/ Outbound Site AM Peak Frontage Road Traffic Percentage Passing Trade Sec. 5.3 N/A

CBD In/ Outbound Site AM Peak Frontage Road Traffic Trip Generation Sec. 5.3 N/A

Intermission

Data Analysis – Discussion of Results

Very low R2 results for influence of:

− Service time on queue length

− Service time on trip generation

− Number of service booths on service times

− GFA on trip generation

Data Analysis – Discussion of Results (Continued)

Frontage Road Site AM Peak Hour in CBD Direction Vs Trip Generation

No clear correlation or relationship can be formed. Similar results and conclusions

drawn for trip generation and CBD bound or two-way frontage road traffic

1

2

34

5

67

89

10

y = 0.0166x + 85.717

R² = 0.11860

50

100

150

200

250

0 500 1000 1500 2000 2500 3000 3500 4000

DC

O G

en

era

ted

Tri

ps

CBD-Bound Traffic Volumes - Site Peak

AM Trip Generation vs CBD-Bound Traffic (Site Peak)


Outlet Gross Floor Area (GFA) relationship to Trip Generation

No correlation between generated trips and GFA of the DCO’s.


Ziper drive-through outlet has a GFA of 7m2


Frontage Road Site AM Peak Hour in CBD Direction and Queue Lengths

View with caution as there are other influencing factors such as accessibility of traffic

from both directions of the road, service times and the number of vehicles served.


Trip Generation Relationship to DCO Queue Lengths

Shows a relationship between queue lengths and trip generation, however other

contributing factors that influence trip generation as a dependent variable

12

3

4

5

6

7

8

9

10

y = 0.0402x + 2.4677

R² = 0.6679

0

2

4

6

8

10

12

14

0 50 100 150 200 250

Qu

eu

e L

en

gth

(V

eh

)

Site AM Peak Trip Generation

Queue Length Relationship to Trips


Staff Number Impact on Service Times

Suggests that a higher number of staff results in an increased service time. Intuitively not

logical. More staff to handle the peak, but service times increase as business increases.

Nature of the relationship rather than dependence.

1, 2, 3

4

5,

6

7

8

9

10

y = 0.8746x + 1.2898

R² = 0.6436

0

1

2

3

4

5

6

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Ave

rage

Ser

vice

Tim

e (m

in)

Number of Staff

Number of Staff to Service Time


Staff Number Impact on Trip Generation

Results probably indicate correlation rather than dependency.

1

2

34

5

67

89

10

y = 39.943x - 12.431

R² = 0.31390

50

100

150

200

250

0 1 2 3 4 5Trip

Ge

ne

rati

on

AM

(Sit

e P

eak

)

Number of Staff

Number of Staff to Trip Generation


Service Booth Numbers Impact on Trip Generation

Higher number of service points are operated by outlets to cater for the business’s

generated trips. Therefore, the relationship is probably more correlation than dependency

12

34

5

67

89

10

y = 37.517x - 14.655R² = 0.6149

0

50

100

150

200

250

0 1 2 3 4 5 6 7

Trip

Ge

ne

rati

on

AM

(Sit

e P

eak

)

Number of Service Booths (Ordering + Payment + Collection)

Number of Service Booths (Total) to Trip Generation


DCO Location Relationship with CBD Inbound Vs Outbound Traffic

Determine possible relationships between the accessibility of each DCO location to

capture customers from CBD inbound and CBD outbound traffic

Reasonable expectation that the location of DCO’s that were best suited to capture

the AM CBD inbound traffic would attract higher trip generation rates

Analysis however, showed no distinct differences in the average DCO’s trip

generation or passing trips based on location

Conclusions

Significantly more trips generated in the AM peak than PM peak

Based on six-day surveys, very low number of customers on Saturday and most

outlets closed on Sunday

Based customer interviews there is a high proportion of passing trips throughout

the day (average 83%) also verified by postcode data

Inter-relationships identified in Table 1, whilst indicative of some dependence, can

be explained by reasoning of normal operations of a business such as DCOs

Some correlation between road frontage traffic volumes and trip generation,

however the R2 relationship is not statistically significant

Does not appear to be a correlation of GFA to trip generation

Appears to be some correlation between trip generation and queue lengths

Conclusions ( Continued)

Outlet management confirm that the number of staff serving is increased during site

peak times to reduce service times, also designed to manage queue lengths

Service times across all outlets generally consistent, with a range of 2:41(min:sec)

to 5:29 and average of 3:53. A “levelling out” of customers an outlet can serve

based on the coffee making equipment they have?

Maximum queue lengths:

− Ranged from 2 to 11

− One maximum queue of 2, two maximum queue of 11

− Remaining seven maximum queue was between 5 and 7

− Overall average maximum for all outlets of 6.7 vehicles

− Queuing capacity of all sites sufficient to avoid queued vehicles onto roadway

− Customers’ limited tolerance to waiting times?

Conclusions – Other influencing factors

Visible exposure to passing traffic

Ease of access to the site

Ease of site egress

Quality and visibility of signage and advertising

Reputation, quality of coffee, food and service

Type of coffee machines used and capacity to produce a maximum rate of coffees

Recommendations

With the exception of a small number of outlets surveyed, due to local circumstances

and excluded as “outliers”, a range of trip generation rates could be reasonably

adopted between 70 and 130 AM peak hour trips

Recommendations (Continued)

Range of values between 70 and 130 trips in the AM peak hour be adopted as a

baseline estimate

The average trip generation for the AM site peak calculated for all DCOs of 105

falls within this range

When assessing proposed DCO developments, selection of an appropriate traffic

generation rate should consider the range of variable influencing factors

Recommended that the average passing trip percentage of 83%

What rates to use for Traffic Impact Assessments?

Baseline range 70 to 130 trips

Whilst R2 not significant there are still evident relationships:

− Frontage road traffic

− Visible exposure to passing traffic

− Ease of access to the site

− Potential customer catchment

Other factors may be unknown at Development Application stage, such as:

− Quality and visibility of signage and advertising

− Reputation and quality of coffee, food and service

− Number of service booths, staff and coffee making capacity

− Seating

What rates to use for Traffic Impact Assessments? (Cont)

Be careful about road frontage traffic and trip generation assumption

This outlet captures a large industrial access restricted area

AM Peak traffic 68 vehicles generating 88 trips (44 vehicles)

What rates to use for Traffic Impact Assessments? (Cont)

Summary of key traffic impact considerations

Baseline trip generation rate of 70 – 130 peak AM trips

Exposure to frontage road traffic

Consider capture of CBD bound traffic in AM

Passing trade – 83 %

Likely maximum queue lengths – Average maximum approximately 7, maximum 11

Visible exposure to passing traffic

Ease of access to the site

Ease of site egress

For proposed sites with seating use parking rates for cafe

Any other known influences such as proposed number of service booths

Acknowledgements

Bitzios Consulting would like to acknowledge

− Vince Taranto, RMS Leader Road Network Analysis for management, support

and assistance throughout this study;

− Traffic Data and Control for the extensive traffic and outlet survey work; and

− Drive-through coffee outlets for their cooperation and assistance:

Fastlane Coffee 1, Dubbo NSW Coffee Club, Tingalpa, QLD

Fastlane Coffee 2, Dubbo NSW Di Bella, Bowen Hills, QLD

Starbucks, Mt Druitt, NSW Espresso Lane, Labrador, QLD

Ziper, Concord, NSW The Brew, Bathurst, NSW

Johnny Bean Good, Bathurst, NSW Tico’s Drive Thru, Brooklyn, VIC

trip generation study of drive-through coffee outlets

Education