NAME
WWW::Analytics::MultiTouch - Multi-touch web analytics, using Google Analytics
SYNOPSIS
use WWW::Analytics::MultiTouch;
# Simple, all-in-one approach
WWW::Analytics::MultiTouch->process(user => $username,
pass => $password,
id => $analytics_id,
start_date => '2010-01-01',
end_date => '2010-02-01',
filename => 'report.xls');
# Or step by step
my $mt = WWW::Analytics::MultiTouch->new(user => $username,
pass => $password,
id => $analytics_id);
$mt->get_data(start_date => '2010-01-01',
end_date => '2010-02-01');
$mt->summarise(window_length => 45);
$mt->report(filename => 'report-45day.xls');
$mt->summarise(window_length => 30);
$mt->report(filename => 'report-30day.xls');DESCRIPTION
This module provides reporting for multi-touch web analytics, as described at http://www.multitouchanalytics.com.
Unlike typical last-session attribution web analytics, multi-touch gives insight into all of the various marketing channels to which a visitor is exposed before finally making the decision to buy.
Multi-touch analytics uses a javascript library to send information from a web user's browser to Google Analytics for raw data collection; this module uses the Google Analytics API to collate the data and then summarises it in a spreadsheet, showing (for example):
Summary of marketing channels and number of transactions to which each channel had some contribution (sum of transactions > total transactions)
Summary of channels and fair attribution of transactions (sum of transactions = total transactions)
First touch, last touch, fifty-fifty first/last touch, and even attribution of transactions.
Overlap analysis
Transaction/touch distribution
List of each transaction and the contributing channels
BASIC USAGE
process
WWW::Analytics::MultiTouch->process(%options)The process() function integrates all of the steps required to generate a report into one, i.e. it creates a WWW::Analytics::MultiTouch object, fetches data from the Google Analytics API, summarises the data and generates a report.
Options available are all of the options for new, get_data, summarise and report. Minimum options are user, pass, id, and typically start_date, end_date and filename.
Typically the most time consuming part of the process is fetching the data from Google. The process() function is suitable if only one set of parameters is to be used for the reports; to generate multiple reports using, for example, different attribution windows, it is more efficient to use the full API to fetch the data once and then run all the needed reports.
METHODS
new
my $mt = WWW::Analytics::MultiTouch->new(%options)Creates a new WWW::Analytics::MultiTouch object.
Options are:
user, pass, id
These are the Google Analytics username, password and reporting ID respectively. These parameters are mandatory.
event_category
The name of the event category used in Google Analytics to store multi-touch data. Defaults to 'multitouch' and only needs to be changed if the equivalent variable in the associated javascript library has been customised.
fieldsep, recsep
Field and record separators for stored multi-touch data. These default to '!' and '*' respectively and only need to be changed if the equivalent variables in the associated javascript library has been customised.
patsep
The pattern separator for turning source, medium and subcategory information into a "channel" identifier. See the
channel_patternoption under summarise for more information. Defaults to '-'.channel_map
This is a hashref of channel name (after applying
channel_pattern) that maps the extracted name to a more friendly name. For example, if channel_pattern is 'med-subcat', then direct traffic appears as '(none)-(none), organic traffic as organic-(none), etc. An appropriate channel_map might be:channel_map => { '(none)-(none)' => 'Direct', 'organic-(none)' => 'Organic' }date_format, time_format
The format to be used for printing dates and times, respectively, using strftime patterns. See "strftime Patterns" in DateTime for details. Defaults are '%d %b %Y' (e.g. 1 Jan 2010) and '%Y-%m-%d %H:%M:%S' (e.g. 2010-01-01 01:00:00).
ga_timezone, report_timezone
Timezone used by Google Analytics, and timezone to be used in the reports, respectively. May be specified either as an Olson DB time zone name ("America/Chicago", "UTC") or an offset string ("+0600"). Default is UTC for both.
revenue_scale
Scaling factor for revenue amounts. Useful if, for example, you wish to display revenue in thousands of dollars instead of dollars.
debug
Enable debug output.
get_data
$mt->get_data(%options)Get data via the Google Analytics API.
Options are:
start_date, end_date
Start and end dates respectively. The total interval includes both start and end dates. Date format is YYYY-MM-DD or YYYYMMDD. (These dates are with respect to the report timezone).
summarise
$mt->summarise(%options)Summarise data.
Options are:
window_length
The analysis window length, in days. Only touches this many days prior to any given order will be included in the analysis.
single_order_model
If set, any touch is counted only once, toward the next order only; subsequent repeat orders do not include touches prior to the initial order.
channel_pattern
Each "channel" is derived from the Google source (source), Google medium (med) and a subcategory (subcat) field that can be set in the javascript calls, joined using the pattern separator patsep (defined in new, default '-').
For example, the source might be 'yahoo' or 'google' and the medium 'organic' or 'cpc'. To see a report on channels yahoo-organic, google-organic, google-cpc etc, the channel pattern would be 'source-med'. To see the report just at the search engine level, channel pattern would be 'source', and to see the report just at the medium level, the channel pattern would be 'med'.
Arbitrary ordering is permissible, e.g. med-subcat-source.
The default channel pattern is 'source-med-subcat'.
channel
A hashref containing channel-specific options. This is a mapping from channel name (the friendly name, given in the channel_map) to option hash.
Currently the only option is 'requires_first_touch' (boolean). If set, a transaction will only be attributed to the channel if it received the first touch in the analysis window. This is mainly used to correct for over-attribution to the direct channel. Example:
{ Direct => { requires_first_touch => 1 } }adjustments
A hashref containing transactions and revenue corrections that may be applied for a given day. This allows, for example, compensation for lost data for short periods of time. A typical form of the adjustments hash is:
{ 2010-09-01 => { transactions => 1.4, revenue => 1.3 } }which would apply correction factors of 1.4 and 1.3 for transaction counts and revenue respectively for any transactions occurring on the date 2010-09-01.
report
$mt->report(%options)Generate reports.
Report Type Options
all_touches_report
If set, the generated report includes the all-touches report; enabled by default. The all-touches report shows, for each channel, the total number of transactions and the total revenue amount in which that channel played a role. Since multiple channels may have contributed to each transaction, the total of all transactions across all channels will exceed the actual number of transactions.
even_touches_report
If set, the generated report includes the even-touches report; enabled by default. The even-touches report shows, for each channel, a number of transactions and revenue amount evenly distributed between the participating channels. For example, if Channel A has 3 touches and Channel B 2 touches, half of the revenue/transactions will be allocated to Channel A and half to Channel B. Since each individual transaction is evenly distributed across the contributing channels, the total of all transactions (revenue) across all channels will equal the actual number of transactions (revenue).
distributed_touches_report
If set, the generated report includes the distributed-touches report; enabled by default. The distributed-touches report shows, for each channel, a number of transactions and revenue amount in proportion to the number of touches for that channel. Since each individual transaction is distributed across the contributing channels, the total of all transactions (revenue) across all channels will equal the actual number of transactions (revenue).
first_touch_report
If set, the generated report includes the first-touch report; enabled by default. The first-touch report allocates transactions and revenue to the channel that received the first touch within the analysis window.
last_touch_report
If set, the generated report includes the last-touch report; enabled by default. The last-touch report allocates transactions and revenue to the channel that received the last touch prior to the transaction.
fifty_fifty_report
If set, the generated report includes the fifty-fifty report; enabled by default. The fifty-fifty report allocates transactions and revenue equally between first touch and last touch contributors.
transactions_report
If set, the generated report includes the transactions report; enabled by default. The transactions report lists each transaction and the channels that contributed to it.
touchlist_report
If set, the generated report includes the touchlist report; enabled by default. The touchlist report lists the touches for each transaction in chronological order. Note that this can be a very large amount of data compared to other reports.
transaction_distribution_report
If set, the generated report includes the transaction distribution report; enabled by default. The transaction distribution report shows the number of transactions that had one touch, two touches, etc, both by channel and as a total.
channel_overlap_report
If set, the generated report includes the channel overlap report; enabled by default. The channel overlap report shows the number of transactions that were touched by 1 channel, 2 channels, etc, and the number of transactions by channel combination.
Report Output Options
filename
Name of file in which to save reports. If not specified, output is sent to STDOUT. The filename extension, if given, is used to determine the file format, which can be xls, csv or txt.
format
May be set to xls, csv or txt to specify Excel, CSV and Text format output respectively. The filename extension takes precedence over this parameter.
Report Formatting Options
title
Title to insert into reports.
column_heading_format
The cell format (see "CELL FORMATS" in WWW::Analytics::MultiTouch) to use for column headings.
column_formats
An array of one or more cell formats (see "CELL FORMATS" in WWW::Analytics::MultiTouch) to use in a round-robin manner across the columns of the data.
header_layout, footer_layout
Page headers and footers. See "HEADERS AND FOOTERS" in WWW::Analytics::MultiTouch for details.
strict_integer_values
If set, transactions and revenue will be reported in integer formats. Where a reasonable number of transactions are being counted, the fractional part of the transaction count in a distributed transactions report is rarely of consequence, and for some the concept of a fractional transaction attribution can be a distraction from the key messages of these reports, so this option helps to keep it simple.
heading_map
A mapping from default report headings to custom report headings. For example,
heading_map => { Transactions => 'Distributed Transactions', 'Revenue' => 'Distributed Revenue (US$)' }
For every type of report, (all_touches_report, first_touch_report, transactions_report, etc), report-specific formatting options can be given in a hashref with corresponding name, e.g. 'all_touches', 'first_touch', 'transactions'. For example,
column_heading_format => {
bold => 1,
color => 'white',
bg_color => 'gray',
right => 'white',
},
column_format => [
{ bg_color => '#D0D0D0', },
{ bg_color => '#E8E8E8', },
],
all_touches => {
column_heading_format => {
color => 'blue'
}
}
The report-specific options are merged with the top level options and then used.
all_touches_report
even_touches_report
distributed_touches_report
first_touch_report
last_touch_report
fifty_fifty_report
touchlist_report
transaction_distribution_report
channel_overlap_report
These implement the individual reports, taking options similar to those described under report above.
DEVELOPER METHODS
As well as the user API methods list above, there are also a number of methods that have been exposed as part of the API for developer purposes, e.g. for developing subclasses to override specific functionality, or for integrating into systems other than Google Analytics.
set_data
$mt->set_data(start_date => 20100101,
end_date => 20100130,
transactions => \%transactions,
);Instead of invoking get_data to retrieve data from Google Analytics, it is also possible to set data directly - e.g. data collected through another mechanism, or data from Google Analytics that has been saved to file. set_data allows you to directly specify the data for subsequent analysis.
Parameters 'start_date' and 'end_date' are used for reporting and should have the form YYYYMMDD.
Parameter 'transactions' is a hash of transaction ID to [date (YMD), list of touches] (see split_events for description).
split_events
@events = $mt->split_events($cookie_value)Splits event label (i.e. 'multitouch' cookie value) into a list comprising order and touch arrayrefs. A touch has format [ source, medium, subcat, time ] and an order has format [ '__ORD', transactionID, revenue, time ].
condition_entry
($conditioned_key, $conditioned_touches) = $self->condition_entry($key, \@touches)condition_entry is called by get_data for each data entry retrieved from Google Analytics. It is possibly useful for subclasses to override, in case any special data conditioning is required. $key is the event label (transaction ID) and @touches is the list of touches, each touch being an array as described under split_events. Conditioning might include removal of duplicates, or normalisation of transaction IDs.
parse_config
$opts = WWW::Analytics::MultiTouch->parse_config($opts, $config_file)Parses $config_file and merges options with $opts.
RELATED INFORMATION
See http://www.multitouchanalytics.com for further details.
AUTHOR
Jon Schutz, <jon at jschutz.net>
BUGS
Please report any bugs or feature requests to bug-www-analytics-multitouch at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=WWW-Analytics-MultiTouch. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.
SUPPORT
You can find documentation for this module with the perldoc command.
perldoc WWW::Analytics::MultiTouchYou can also look for information at:
RT: CPAN's request tracker
http://rt.cpan.org/NoAuth/Bugs.html?Dist=WWW-Analytics-MultiTouch
AnnoCPAN: Annotated CPAN documentation
CPAN Ratings
Search CPAN
COPYRIGHT & LICENSE
Copyright 2010 YourAmigo Ltd.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.