You can add, remove and update user rights by clicking "User management" in the top bar of the giosg console.
Choose "Users" and you can see a list of all users.
Click "+New user" to create a new user, and fill in the email address of the user. This sends a registration link to that email. The link is valid for 48h.
You can define the user rights for each user and edit their information by opening the information of that user. You can also delete users here, and set them offline in the chat if needed.
The chat automatically appears on all pages in your domain (where the giosg script is added), unless you change the settings. The chat is visible only when someone is online in the chat. If you want to limit the chat to certain pages, you need to disable the default settings first. After that you can use rules to specify where the chat is shown.
To disable the default settings go to:
1. Settings -> Domains -> speech bubble icon -> Enable. Change this from Yes to No
2. Settings -> Domains -> cogwheel icon -> Auto suggest chat to visitor. Change this from Always to Only with rules (in the Chat settings section)
After this you can create rules for showing the chat. Use the "Enable chat button" action for showing the chat button when the chat is online. If you want to send proactive autosuggest messages, use the "Send autosuggest" action. Use for example "Visitor's URL as a condition in the rule, to make a rule for a specific page.
First make sure that the giosg script tag is added to your web-page and is loading correctly, and that the domain is added to your giosg account. You can find the script tag in your giosg account by clicking Settings -> Company -> Script tag
The chat elements appear on the web-page once a chat agent is online (have pressed Start).
In order for the chat to work on a website, the domain needs to be added to your giosg account. You can add and manage domains by going to Settings -> Domains. From here you can also access the design settings of the chat by clicking on the speech bubble icon (chat button settings) or cogwheel icon (chat window settings).
When adding new domains, please note that the domain should be in basic form, for example giosg.com instead of www.giosg.com or www.giosg.com/something
In the giosg console, click Settings -> Chat windows to create new chat windows and edit existing ones.
You can edit the texts and position, as well as add your logo to the chat window. If needed you can make changes in the design of the window, by clicking "Open style editor". Here you can edit the css and, if required, override HTML templates.
The chat window that is used in a domain is determined from Settings -> Rooms -> Choose domain -> Chat window (in the Chat settings section)
To edit the chat button, go to Settings -> Domains -> speech bubble icon. Here you can edit the appearance, text and position of the chat button. Change the text on the button simply by clicking on the current text, and type in the new text. The design can be edited by clicking "Design".
The chat button will appear on all pages by default. If you want to decide on which pages the chat button is showing, please use giosg Rules.
There are two different ways to customise autosuggest messages. The first is through the Rooms tool feature and the second way is by using the Rules tool. Using the latter method will give you more possibilities to customise the autosuggest message. Here is how to edit the autosuggest messages:
1) Using the Rules tool, go from the giosg-console to:
-Create new rule
-Add new condition (a condition is what an online visitor has to do on your site in order for an action to trigger, there are many different ones giving you the opportunity to customize your actions, in this case, the autosuggest message)
-Select action to perform: Autosuggest chat
-Type custom autosuggest message
(Check out here Rules instructions)
2) If you don't have giosg Rules, you can edit autosuggest message using the Rooms tool, go from the giosg-console to:
-Choose the correct room
-Choose from the chat settings when you want the autosuggest messages to be sent and finally define the message
Shortly, the methods of calculation we use in our analytics differ from other services, such as Google analytics.
The differences are mainly caused by the way we define the length of the session and by the security settings of the visitors' browsers. Differences can also be a result of delimitations made in other analytics service.
The visitors are recognized based on the cookie information from the browser and the location is recognized based on the visitor’s IP-address.
Mobile chat-window opens in the same page tab, similarly to the desktop version. You can edit chat button colour, agent message box colour and to add a logo on the window header.
-Click “Settings” -> “Rooms” -> Select room to edit
-Click on “Create brand” to create a new layout or “Edit brand to ”
-Select the colours for chat button and agent message box and add a logo to the header
-After your layout is ready, select mobile chat-window language under room settings
You can delete sensitive visitor data such as chat messages, visitor details and lead form content. Benchmark data will be kept for reporting purposes.
This feature lets you choose after how many days data is deleted, any older data will also be deleted once the deletion is activated. As a default sensitive visitor data is deleted every 5 years. The deletion is permanent and giosg will not be able to restore this information.
Select data deletion timeframe:
-Click “Settings” -> “Rooms” -> Select room
-Scroll down to “Delete sensitive data”
-Under “Delete sensitive chat data after x days” type in the amount of days after which all sensitive data will be deleted automatically
-Scroll down and save the changes
A custom room is a way of dividing a domain into "rooms", which enable you to have different teams taking care of different parts of your site. For example, divide your site into Sales, Customer Service and Tech Support, and have these teams take care of chats related to their roles.
Create a custom room :
-> Settings -> Rooms -> Click "Create new room" under "custom rooms"
-> Find and click the new room -> Add the domain for this custom room under “Connected domains” section
Create a Rule that guides the traffic to that specific room:
-> Settings -> Rules -> Add a rule under your main/connected domain
-> Use Conditions to match your Room with the visitors needs, (i.e. visitors URL matches "support"). Add the action "Join room" and select the new room created for the Support team.
Read more about creating custom rooms from our user guide Creating custom rooms
To export data from giosg, complete the following steps from the chat console:
-> Settings -> Reporting -> Custom Reports
-> From the top, select the timespan and which rooms you want to export statistics from (you can select multiple rooms at once).
-> Select which type of report you want from the left (Chat logs, Total Statistics, Total Statistics Hourly, Leads)
-> Select the relevant fields to export from the column in the middle
-> From the right, choose “Download report now” and select the right file format (XLSX, CSV, TXT)
Read more about reports from the giosg REPORTING guide
Note, it's also possible to use our public APIs for reporting.
Instead of downloading custom reports manually, you can automate the reports to be downloaded and sent to one or more emails. You can schedule these reports daily, weekly or monthly.
To schedule your reports:
-> Settings -> Reporting -> Custom Reports -> Select the right room, timespan, report type and fields, (read more on this above Exporting analytic data from giosg)
-> Choose "subscribe to email report" and choose daily, weekly or monthly
-> Click on "Subscribe" to receive the custom report to your mail
How to edit your custom reports and key statistics reports:
-Click on “Settings"
-Click on “Company”
-Click on “Email report”
Edit your custom reports and/or key statistics reports in this view
-Subscribe or unsubscribe to reports
-Create custom and key statistics reports to come (daily, weekly or monthly)
You can find a summary of chat statistics per agent for a selected time range under:
Reporting -> Operator -> Chat statistics Beta
Then select time range and the rooms you want the report to contain. You can select a time range starting earliest from 1.5.2018, older statistics can be found from "Custom reports" or "Chat statistics". As a default the report shows the data for all rooms in the organization.
You can access the chat logs for a specific agent by clicking on the agent's name from the list
How do I interpret the Target report?
Each Target configuration is displayed on its own row. The same configuration can be divided into several experiments.
When you first set up a new Target configuration, the first step is the data collection mode. In this mode we collect data for the algorithm by having only random and control groups. The control group will not receive the treatments at all and the members in the random group will receive the treatment randomly. The treatment is the action you have configured when creating the Target rule.
When we have enough data, the system will go into target mode. In this mode we have three groups
Control and random groups work as before and are there for the comparison. The target group is used to serve the treatment based on our algorithm. The algorithm predicts conversion probabilities and directs treatments to the ones fulfilling the set criteria.
The summary table shows the metrics for the Target group. The same metrics can be found in the results table.
Shows the metrics for each group in the experiment
Shows the groups that are part of the experiment.
Shows the amount of distinct visitors in each group. Each visitor is counted only once in the selected timeframe.
|Treatments||Total number of treatments targeted in this group. A single visitor may have received multiple treatments.|
|Goals||Total number of reached goals by the visitors in this group. A single visitor may have reached the goal multiple times.|
|Conversion %||The conversion rate of the group.|
|P( x > reference group) (Test for statistical significance)||
Describes the probability that the group performs better than the reference group. >50%: The performance is better than the control group <50%: The performance is worse than the control group Results above 95% and below 5% are considered statistically significant. At the beginning of the experiment the values can vary a lot.
Describes the range where the conversion rate is for each group with 95% probability. We model the conversion rate p with beta distribution. The 95% confidence interval is the most narrow range that captures 95% of the probability mass. This is also known as the highest posterior density or highest density interval.
The values for the selected period are calculated from cumulative weekly counts of reached goals and unique visitors. E.g. if the selected period is four weeks, the first week’s values come from the first week’s counts. The second week’s values are calculated from the counts for first AND second week, third week from first, second and third weeks’ counts etc. Note that Unique Visitor Count for two weeks is usually not the first and second weeks’ counts summed, because there most probably are visitors that visit the site on both weeks.
Shows how many visitors were in each group per week. These are unique visitors per weekly period, so one visitor is counted only once per week.
You can monitor tags from reporting section: from real time section, operator stats or custom reports. In operator stats you can check which tags have been added to a certain conversation, whereas from custom reports you will also get a summary of total tag usage within the selected time period. In the real time report you will see a summary of tags for the day.
Start by going to reports section, then select either:
-select a chat agent whose chats you wish to view
-chats on the list with contain at least one tag
-click on the timestamp of a conversation to see the tags added to that chat.
-select chat logs (under "Available reports"-menu)
-select tags (under "Choose fields to export"-menu)
Target enables you to trigger treatments such as reductions or contact forms to visitors that really need it. The triggering is based on our machine learning algorithms. By automatically analysing the visitor behaviour, we can target treatments more effectively.
Target always needs a goal to be able to function. You can create a new goal by going to Settings-> Goals-> Create a new goal
Set the parameters for the new goal
After you are happy with all the settings, make sure to save the goal
After creating the goal, you have to create a new rule which uses Target. Open Settings->Rules and click “Create new rule” next to the room you want to enable Target
You should see the following view
To enable Target to this rule, click “Activate Target”-button. The rule doesn’t have any other conditions and the rule is triggered to the visitor only when the algorithm determines the visitor should be targeted. The view changes a bit to show Target-specific information
After enabling the Target, set up the action you want to do to the visitor. This can be chat suggestion or discount banner or something else that provides the visitor some extra value. You can also add multiple actions if you want.
Select the goal you want the algorithm to use. This is the goal that the algorithm aims and runs the rule actions only to the potential visitors.
When the rule is ready, save it. After saving the rule the following steps will happen
In the data collection mode the visitors are divided to two different groups
Control group is used to see the conversion without the treatment that is set in the Target rule. On the other hand, visitors on the random group will receive the treatment randomly. This is to collect the behavioral data from visitors and decide which visitors will receive the goal.
After we have enough data for the algorithm, we create a prediction model. When the model is ready, it will be taken into use automatically. The website visitors are now divided to three groups instead of two in data collection mode. The groups are
The first two act as before, but Target group will receive the treatment based on the prediction model. If the model sees the visitor as potential it will notify our system and the visitor will receive the treatment.