How unit 2 is examined
This unit covers social network concepts, how web data is captured, outcome and competitive data, analytics team structure, and data quality issues. No topic was asked in the supplied papers, so all are short.
The Social Networks Perspective - Nodes, Ties and Influencers
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>The social networks perspective studies people or organisations as nodes joined by ties, and explains behaviour through the pattern of those relationships.</mark>
Key points.
- A node is an actor such as a user, page or company, and a tie is a relationship between two nodes such as friendship, follow or retweet.
- Ties can be directed or undirected, and strong or weak, so a follow is directed while a friendship is mutual.
- An influencer is a node with many ties or central position, whose posts reach and sway many other nodes.
- Influence is measured by degree (number of ties), so a node with more followers is usually more central.
Social Network, Web Data and Methods
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Social network data is relational data about who is connected to whom, collected from the web and analysed with network and web-analytics methods.</mark>
Key points.
- Web data comes from platform APIs, page crawling, server logs and user-generated content such as posts, likes and comments.
- Network methods represent the data as a graph of nodes and edges, then compute measures such as degree and centrality.
- Web methods count visits, page views and conversions to show how users behave on a site.
- Both methods are combined so that who a user is connected to explains what they click or share.
Capturing Data: Web Logs, Web Beacons, Java Script Tags, Packet Sniffing
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Data capture is the way visitor activity is recorded, using web logs, web beacons, JavaScript tags or packet sniffing.</mark>
Key points.
- Web logs are files the web server writes for every request, holding IP address, time, page and status code, so they need no page change but miss cached pages and bots are mixed in.
- A web beacon is a tiny 1x1 transparent image whose request tells the server that a page or email was viewed.
- A JavaScript tag is code placed on each page that runs in the browser and sends visitor data to the analytics server, and it is the most common method today.
- Packet sniffing reads network packets passing between visitor and server, so it needs no page tagging but needs special hardware and cannot see encrypted content.
Outcome Data: E-commerce, Lead Generation, Brand/ Advocacy and Support
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Outcome data measures what a website achieves for the business, such as sales, leads, brand engagement and support resolution, rather than only how much traffic it gets.</mark>
Key points.
- E-commerce outcomes are orders, revenue, average order value and conversion rate.
- Lead generation outcomes are form submissions, sign-ups, downloads and enquiries that sales can follow up.
- Brand and advocacy outcomes are shares, mentions, sentiment and reviews that show customers promoting the brand.
- Support outcomes are queries resolved online, call deflection and satisfaction ratings.
Competitive Data: Panel Based Measurement, ISP Based Measurement, Search Engine Data
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Competitive data is information about rival sites and the whole market, gathered from outside your own website.</mark>
Key points.
- Panel based measurement installs software on the computers of a sample of volunteers and projects their browsing to the whole population, so it is only as good as the panel is representative.
- ISP based measurement uses anonymised traffic records from internet service providers, which cover far more users than a panel.
- Search engine data shows which keywords and phrases people search and how rivals rank and advertise for them.
- Competitive data is an estimate, so use it to compare trends and share, not exact counts.
Organisational Structure
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Organisational structure is how the analytics team is placed and governed in a company so that its insights reach decision makers.</mark>
Key points.
- In a centralised model one team serves the whole company, giving consistent data and standards but slower response to each unit.
- In a decentralised model each business unit has its own analysts, giving speed and context but inconsistent metrics.
- A hybrid or centre-of-excellence model keeps standards and tools central while analysts sit inside business units.
- Whatever the model, there must be executive support, clear ownership of data and analysts who can turn numbers into actions.
Type and Size of Data, Identifying Unique page Definition, Cookies, Link Coding Issues
<span style="display:inline-block;padding:.16em .6em;border:1.5px solid currentColor;border-radius:999px;font-size:.68em;font-weight:700;letter-spacing:.06em;text-transform:uppercase;opacity:.75">Not asked since 2022</span>
Definition. <mark>Data quality depends on the type and size of data collected, on defining each unique page correctly, on cookies used to identify visitors, and on how links are coded.</mark>
Key points.
- Data is structured (tables, counts) or unstructured (text, images), and web data volume is large, so it is often sampled or summarised.
- A unique page must be defined by its content, because dynamic URLs with session IDs or parameters can make one page look like many, and the reverse.
- Cookies are small files that identify a browser and separate new from returning visitors, but they are lost when deleted or blocked and count one person on two devices as two visitors.
- Link coding issues arise when campaign tracking parameters or redirects are missing or wrong, so traffic is credited to the wrong source.
Last-minute revision
- A node is an actor and a tie is a relationship; influencers are highly connected nodes.
- Degree is the number of ties of a node.
- Social network data is relational and is analysed as a graph.
- Web logs are server-side records of every request.
- A web beacon is a 1x1 transparent image used to count views.
- A JavaScript tag runs in the browser and sends data to the analytics server.
- Packet sniffing reads network packets and cannot read encrypted content.
- Outcome data covers e-commerce, leads, brand advocacy and support.
- Panel data is sample based, ISP data is broad, search data shows keywords.
- Organisation models are centralised, decentralised and hybrid.
- Cookies identify browsers, not people.
Memory hooks
- Capture methods: "Log, Beacon, Tag, Sniff" from server file to pixel to script to packet.
- Competitive data: "Panel is a Sample, ISP is Broad, Search is Keywords".
- Org models: "One team, Many teams, Mixed team".
- Cookie counts a device, not a human.
Coverage checklist
- The Social Networks Perspective - Nodes, Ties and Influencers: no past questions.
- Social Network, Web Data and Methods: no past questions.
- Capturing Data: Web Logs, Web Beacons, Java Script Tags, Packet Sniffing: no past questions.
- Outcome Data: E-commerce, Lead Generation, Brand/ Advocacy and Support: no past questions.
- Competitive Data: Panel Based Measurement, ISP Based Measurement, Search Engine Data: no past questions.
- Organisational Structure: no past questions.
- Type and Size of Data, Identifying Unique page Definition, Cookies, Link Coding Issues: no past questions.