Importance of modelling

This morning reading the article that a report commissioned by Google found that the internet economy is worth £100 billion I couldn’t help but have a little giggle to myself: As someone who has been preaching and preaching the need to monitrise online activity I welcomed this high level demonstration of grabbing headlines with big figures, but also couldn’t help but smirk at the problems this report would get in particular web departments if it used to value their activity.

Two problems would be raised:


The report is obviously a massive PR activity by Google:
Report shows how much business goes on online = business rush out to get bigger online presence = bigger $ ad revenue for the world biggest online ad firm.

I can imagine the arguments as to whether such activity was correct and justified by web teams and whether, rightly so, did it invalidate the whole argument.

Methodology flaws:

Second problem would be the massive jumps and estimations that were required to reach this figure.  In creating this £100 billion throughout the document figures had to be estimated or a particular ‘best method’ had to be used just to get to the figure .  I don’t want to get technical, but the below should tell its own story:

“We estimate the value of goods and services researched online but purchased offline at £40 billion in 2008” (p12 of report)

We all know that getting to the above figure needs a hell of a lot of estimations. So you could, if so inclined, flaw the whole thing picking at its method of estimation and again, rightly so.


Then you miss the whole point. This report serves a purpose of putting an activity that by its nature is a vague proposition (internet use) into a context that everyone understanding (£ signs). And if Google the company with more data than any other in history has to make ‘estimations’ and ask the authors to paint with broad brush strokes, than individual web team should not be afraid to do so if it helps all externally to understand better and value what you do.

As a final thought, I also like this report and will refer to it again as it a useful example of how sometimes you need to forget the myriad of available statistics to make a point. Internet reporting often reaches a point of inertia through having too much data. Those working in the web will not make the jumps and leaps that the offline world has to make because of its lack of information as the online analyst is desperately trying to find the perfect correct figure through their millions of data columns at their disposal.

Online value can only be properly shown by combining with the offline world and that means estimations and sometimes best even best guesses- and if this makes you uneasy just remember every other department in your organisation is estimating and guessing every day!

A Generic Dashboard

Here an example of our work

Marieke Guy – session “Big and Small data”

Marieke starting won. Big and small data


Most of us work in web team- one I management couple of researchers.

What is data? “Stuff” seems to be the best answer.

Going through definitions: Big data is the watch word – shows example of LHC. What is too big data- are institutions getting more data then they need and makes the example that data is growing expotentially- that is the theme of the start

Fact 1 in 9 people in the world are on Facebook – that is a lot othat’s –  so I always wonder why is Facebook  insights so crap?

Mentioned crowd sourcing – both human and machine source.


“openness is coming up constantly ” there are the themes: big data, openess of the data and what can you do with it?

Break up to talk about our data:

Our group:

Decision making is imports are universities asking the right question? And do they have the right answers. Are the people asking the right story

Hawksey makes the point that journalism guardian NYT playing around with open data-  and allowing people to make things up

Marieke says data will effect HE – by requests, REF etc…. But it can help you !


Good point is raised about – can you trust the data – can you trust google’s daTa?

I think it is important that whether you thinkthat the information in key or the human interpretation, my point is that it depends on who the end goal.

Are we asking the right question and that every university has it’s own problems- only you can understand your organisation .

We are talking about open data and problems of data: real sceptism of whether data is comparable, and whether we will compare like with like.

Open Data






new dawn for SiD:

The last three years have been great- when i set up SiD with a laptop and some business cards I could only dream that the company would would flourish as it has- so a big thanks to all round.

However, as we are setting up the UK office I have used this opportunity to have a re-focus on what we are doing- and ultimately ask the question- what exactly we want to achieve?  The answer was to provide a better, relevant product and have well trained,  positive staff who have developed whilst working at the company. In line with this we will be looking to change:

  1. How the company will be run
  2. How we provide our service to our customers

How the company will be run.

On a personal level I consider the most important factor that prevents IT companies providing  a good service  for clients is having staff who not only understand what the client requires, but also have an understanding of technically how it should work- basically rounded people in the industry. In this world of specialism and segmenting, often you have people who are either technical and would never meet (or be allowed!) to meet a client unless sharproned, or a client facing sales person who knows the talk, but would not know the first thing about the nuts and bolts of what they are selling. In my view this leads to the client not getting a good service; often the language used is jargon filled and meaningless leading to service agreements that are tick lists instead of being  driven to providing a meaningful end product.

On the other side of this fence: when employing people you often end up with  either technical efficient people who are really lacking customer empathy or sales people who are lacking in service delivery- and often this has been engrained into them over years of working in IT companies where this distinction is not just in place, but actively pushed.

So we are looking to set up the UK office differently: Each team will consist of a team leader (I shall be the first) and three others who we are planning to take get straight from university/higher education. The key emphasis will be that each staff member will be directly responsible for technical implementation/project management/sales- basically the whole lot. The training will be broad and wide ranging.

Yes, of course, there will be specialisms within the team, but each member of the team is involved in all aspects of the running of the company- i.e. doing client sales pitching as well as technical support. Hopefully this will provide the customer with a better service.

We want to work with people straight from education for a couple of reasons: Firstly, is that there seems to be such a lack of opportunity for clever, motivated people to get a footing in IT in the UK and particularly Scotland. Secondly, we do want people who have not become ingrained with the duality of roles in IT companies and thirdly on a personal level I remember how hard it was after University for someone full of energy and optimism being pushed into narrow job roles which were basic admin which became very boring very quick.

Will we be able to get the right type of person? I will keep you posted!

How we provide our service to our customers

Our previous company statement stated:

“Our underlying ethos is to bend the technology available to fit into your requirements, rather then you having to bend to fit the technology.”

This holds true, but we will be adding a key word to it all: “understanding”: We will be focusing further at understanding where exactly the client is currently and what they want to achieve, and then look at what is the best way to gain the data that will provide the necessary results:  It could be qualitative or quantative, a free product or paid product- it could be tracking or not tracking, but no products or services will be used until we have a clear understanding of the requirements. Once the data is gained it will be analyzed and presented. The next cycle will be to use this analysis and the clients input to set the next achievement and so on, the aim being a full interaction with the client at all times

Sounds simple and obvious? Yes it is, but behind it are two key lessons we have learned over the years: a. Making sure the data is what the customer asked for, therefore making sure it is useful and will provide some insight: It is very tiring and boring producing reports that no-one is interested in and no-one reads. b. Too often products are installed at great cost of time and money and not used at after the first week and just produce the one dimensional automatic reports.  Clients’ effort should be used in telling us what you require in pain english- it is our job to do the rest.

This is all very exciting and as a project has already got good  feedback- I look forward to yours! Please feel free to tell us what you think!

Ranjit Sidhu

Director- SiD


twitter: @rssidhu

By |blog|Comments Off on new dawn for SiD:

Jargon Junkies

Here we will be going through what we see are some of the fundamental problems with online analysis- basically trying to answer the question “with so much information at our finger tips why is online analysis seen as being so crap?” – call it a therapeutic  process

“sorry I don’t understand what is the difference between a page view, visit and visitor?”

How many times in the past have I heard this from a client when they look at a report or at a analytics providers ‘user friendly’ interface? Initially, as so many, you reel of the standard answer of “a person goes to you website and looks at the homepage and then leaves…” answer, however, the client is right- why when talking to a  particular client whose organisation does a very particular objective do we use such jargon?

The page view, visit or visitor is simply the easiest activity analysis can track and has became the standard as it did and still does mirror the requirement of  advertisers who are interested in click throughs and page impressions. Conversion has been tacked on to the list and often now this all encompassing phrase ‘engagement’, but often these great leaps in functionality mean very little to many many businesses, who use it just because they have to without it having any real meaning to them.

The problem is that all these terms generic, meaningful to online analysis in its own realm, but not to the clients own. It often leads to the key frustration on the part of the customer as they desperately try to convey in their terms what they require for the analyst to return with reports with these technical phrases. The hardest part of analysis and the most frightening is to lose the language and be bare to your customer’s actual non jargon requirements (the dreaded why? questions), as than you realise even though you have hundreds of graphs and tables how little useful and relevant information you have from only reports

In my experience, meaningful analysis invariably needs the analyst  to find a few key findings that take time and effort in combining the online results with other information that is offline stuck in their organisation somewhere, this is time consuming, does not find a standard pattern and often requires imaginative thinking to put it all into context.

Perhaps this is the key problem as this requires a completely different mentality and basic skill set and also requires the confession that the shiny new analytics system is absolutely useless if its data is not processed by that antiquated object: The human brain.


This my talk at the IWMW10 also see

Adding value:university

Firstly, thanks for all the great feedback from the IWMW10, it’s been over-whelming! It was great to know so many people are tuned in to what I am trying to communicate: Showing the value of what web teams do in real terms is fundamental to their ability to attain budgets in the near future.

Since last week I have been thinking about what should be next steps, the importance of the timing was further enhanced by the next day listening to Vince Cable talking about funding for the universities and the Lord Browne report that is sure to be challenging. Unlike previous years, there is a real sense of urgency that things need to happen now.

Further, from an outside prospective, I think it is important that institutions are seen to be proactive in this.

With this in mind I have come up with what is important to get in place to help universities follow up from the IWMW10: Here are my suggestions- feedback, as ever, is essential:

1. Creating a structured program for showing value for certain areas of your online site

One point that became clear at the event and subsequently is that although the talk was a good starting point, more detailed examples of providing value would be helpful. It still feel that in many areas the true value of what the online university community is doing is undervalued and perceived as difficult to communicate. I will be looking at giving examples of how to provide a detailed value analysis of:

  • Prospective postgraduate students applications
  • •Mobile Apps
  • •Particular sections of a site (eg Editorial)
  • Providing more detailed scoring of events on the site and its reporting

2. Getting together “meaningful” institutional statistics.


Before I left for Australia in 2006 we were providing Sector Statistics for 22 universities at IWMW 2006 and around 40 before I left. What I am hoping to do is pick that up and carry on with it, but this time actually providing more meaningful reports for the community. All reports will be aggregated data and used basically for:

a. benchmarking so to provide background for the statistics communicated

b. To provide “industry standard” statistics for use when their own institutions statistics are not available.

Example of these reports would be:

  • average monthly applications online
  • average applications per country per month
  • average applications per site section (undergraduate/postgraduate/international)
  • average level of use for mobile application by current students

These reports would obviously start with the basic reports and grow in number and complexity over time. New reports could be sourced from the community as to what is required at differing times.

I will be sending an email out asking organisations to join, but if you could register your interest that would be a helpful start.

3. “Kick Start Day”: A day per institution


I am also mindful that experience has taught me that getting going on something like this is often hard and can get sidelined when the everyday issues come knocking. Often a kick is needed to build up the velocity to get reporting.

With this is mind I am proposing is a “Kick Start Day”. This would be a day per university where we would go through the process and at the end provide a one page report as a template they can follow. We would go through:

  • what they would like to report on
  • how to get the relevant data out
  • Create and a one page report.

I would be providing this at cost service and limited to a one per week on a first come first served basis.


So, this is my plan to get going, as you can imagine it does entail a fair amount of work and commitment from my side so really need to know whether it is worth it. Working with the Universities in the past I know that we can do some amazing things and create new reporting and systems.

What I ideally like to know:

1. Do you think a structured schema of measuring value for parts of your online would be valuable?

2. Would you be provisionally interested in the sector analysis and statistics?

3. Would you be provisionally interested in the Kick Start Day?

4. Is there anything I have missed that you think would be important?

Look forward to your comments and ideas. Leave them here or email or twitter: @rssidhu

HTML Snippets Powered By :