A global white-goods brand, lets call them GlobalBrandlux.com, sent me a survey asking what I thought about my recent website experience. Here’s my response – I wanted to be honest and constructive:

“Sorry, but this is one of the worst product websites I have come across for being able to find (and identify) a particular product. On-site search is a key feature not working that appears to be an after thought. For example…

Returned results for my queries are either irrelevant to me, poorly described (so I don’t even know if it is relevant to my query), or produce zero results. After many, many attempts and double checking that my query for a specific product number exactly matched what was written on my appliance, I discovered the zero results was due to your search function not recognising any “-” characters in the number, or spaces. It has to be a continuous sequence of numbers. That is not how it is written on my appliance and is just plain annoying!

The overwhelming lack of product photographs and the use of superficial descriptions makes matters worse.”

The impact of poor user experience…

I found the following Google Trends graph interesting. It shows the drop in query volume for the GlobalBrandlux.com brand term over the years – a very large 60% drop from their peak of June 2004. May be other people have had similar bad experiences and gone elsewhere… Just a speculation that something is going wrong at GlobalBrandlux.com, not causality…!

So, what happens next – for me the consumer…?

Coupled with the fact that our 5 year old appliances are already starting to fall apart, this web experience made me feel that I will not purchase from this company again (a considerable sum of money). The sad thing is, if the web experience had been a good one – allowing me to order my replacement parts on the first visit – instead of taking me over 5 visits and requiring 6 emails to the customer service department to get it right – they could have gotten away with it. That is, kept me as a future customer.

How many other people go through this bad experience on the web every day….? It amazes me why so many web sites are so poor form a user experience point of view.

Calculating the lost revenue for GlobalBrandlux.com…

• I estimate a household similar to mine(!) renovates its kitchen every 10 years to include a freezer, fridge, microwave, dishwasher, oven/hob, vacuum cleaner.
• I estimate the total cost for a complete set of kitchen appliances from GlobalBrandlux.com at $7,000 USD
• According to compete.com, the US version of the GlobalBrandlux.com website received 39,000 unique visitors in January 2013. This is a company that publishes its sales figures online. So I know that the US market is 26% of their global sales by revenue (2012 figures).
• Assuming this directly correlates with visitor volume, the total global website monthly unique visitors = 150,000
• Lets assume 10% of these are existing customers that would purchase again within 10 years (15,000 visits per month)
• Lets assume half of these people – only 5% of the total traffic – like me, are so p*ssed off with the web experience they decide not to purchase form this brand again. That’s 7,500 people.
• Cost = 7500 x $7000

• = $52.5 million, per month(!) lost revenue

Caveats – there are always caveats with data (make sure you always have yours thought through when presenting your data story)

I deliberated decided to only consider existing customers of GlobalBrandlux.com so that I know I am dealing with people that have previously committed to a significant purchase from them in the past i.e. likely to purchase again – if their experience is a good one.

Apart from the standard issues of data accuracy (multi-device usage, cookie loss etc. – see Chapter 2 of the book, or the my accuracy whitepaper on this subject), here are some of the caveats when making this calculation:

• The number of unique visitors reported by compete.com for Jan 2013 may be a peak (or trough) that can change significantly throughout the year and over time.
• 150,000 unique visitors does not mean that over an extended period, say 12 months, that 1.8m unique visitors came to the website (i.e. 12 x $52.5 million should not be assumed).
• I have assumed each customer would purchase a complete kitchen appliance set. Maybe some would only purchase a microwave, at a much lower value. That said, over a 10 year period, the objective of GlobalBrandlux.com will be to sell an entire suite of kitchen appliances. I have found it quite common that people do this i.e. purchase the same brand throughout their kitchen.

Why are organisations so negligent at this…?

Seriously, some simple JavaScript validation for the on-site search function would go a long way to improving things. By that, I mean removing any non-numeric characters/spaces from a product number that a visitor types in (as it appears all product number sare stored this way at GlobalBrandlux.com). A first-year computer science student can do this, as well as anyone with a basic ability of JavaScript, such as me.

I agree that beyond that, it gets more laborious to improve things – photographing all products (though that has probably been done by someone else in the organisation before), improving product descriptions so that it makes sense to anyone who did not actually build the thing. However laborious these tasks may be, they can surely not come anywhere close to the $52.5 million per month being left on the table.

The point of my simple econometrics calculation is to show that their is a great deal of money being left on the table by organisations not taking the user experience seriously – or even considering it, in the case of GlobalBrandlux.com. Even if my estimations are out by an order of magnitude, it still remains at several million dollars per month!

Have you ever monetised a bad web experience?

© Brian Clifton for Measuring Success, 2013. | Permalink | One comment | Add to del.icio.us
Post tags: cost, site search, usability, user experience

In my analogy:

The organisation is analogous to a website owner/marketer; The road is the web; a car represents a visitor’s browser; The person(s) in the car are the real people who are using the web; Destinations (shops, schools, houses) are the websites. PII = personally identifiable information.

Essentially, as you move down this list the data becomes more personal and therefore privacy becomes more important to the visitor. Also it means your legal obligations wrt privacy increase, as well keeping on top of best practice so that your visitors actually trust you.

Where Does Your Site Fit with Privacy…?

To establish this you need to assess your website by conducting a “tracking audit”. That is, document what tracking methodologies are deployed and assigning each to one of the three classifications above. If any of the data you are collecting is *not* classed as green, ask the following questions of it:

1. Do we need this data?
2. If so, how does it help us optimise our website content or website’s marketing?
3. Is it transparent to the visitor what we are doing with their data?

Often I find that organisation’s inadvertently collect way more information than they actually use, or need. So avoid the privacy hassle and cull any unnecessary data points that infringe upon privacy.

Of course all website owners wish to individualise their data and get personal – because that provides opportunities to tailor content, cross-sell and upwell. There is nothing wrong with that per se, so long as you gain the explicit consent of your visitors first, and stop tracking them if they say no. That is now written into EU Law.

Gaining explicit consent is cumbersome. Raising a red flag to your visitors highlighting privacy is likely to put many of your visitors off. That is not necessary because they fear you are doing something bad. More likely its due to privacy being a complex subject with many ramifications for your visitors that they simply did not contemplate when they decided to visit your site. If in doubt, a person will always err on the side of caution i.e. opt out of your tracking.

This is why I recommend you cull any data collection that is not strictly necessary.

Where Does Google Analytics Fit with Privacy…?

All Google Analytics reports are anonymous and aggregate. That means it fits into the green category. So nothing to worry about, right?

Not quite…

It is possible to break the Terms of Service for Google Analytics and collect PII. I see this often. This can happen inadvertently when visitors receive a confirmation email for a sign-up, or registration etc. That is, the confirmation link in the email includes a clear text version of their email address in the URL. GA tracks URLs by default, so that email address is captured in your reports when the user clicks through on the link.

Solution: Either encrypt the email address in confirmation URL, or use a search & replace filter in your Google Analytics setup to remove it from your reports.

Be Aware of Hidden Tracking Code on Your Site

These days it is rare that the only tracking technology you have on your site is Google Analytics. The plethora of useful third party “widgets” website embed in their site means that pretty much all websites have numerous widgets that provided tracking – either directly to the website owner, or back to the third-party widget owner. Often, organisations are unaware of widget tracking abilities.

If you have any of the following deployed on your site, you are collecting more than just green information and need to assess the privacy impact:

• DoubleClick
• Google Maps
• AdSense
• Disqus
• YouTube
• ShareThis
• LivePerson Chat
• Social plugin buttons (Tweet me, Follow me, Facebook Like, Google Plus, LinkedIn etc.)
• Addthis
• UserVoice

*ALL* of the above set 3rd-party cookies that track individuals (although anonymously).

Useful Tools

There are a number of tools that can help you understand what tracking technologies are deployed on your site. I regularly use the following two (also see my post on The Best Google Analytics Add-Ons):

• WASP (Firefox add-on)
• Google Tag Assistant (Chrome add-on)

The Real Privacy Debate – The Triangulation of Anonymous Information

Related: This recent article from the BBC: http://www.bbc.co.uk/news/technology-21499190

Because the widgets I list above deploy third-party cookies, they have the ability to track visitors around the web – not just on your website. That is, tracking the path of visitors to unrelated websites and what they do there. This is possible because of the ubiquitous nature of these plugins.

For example, from my search history Google can easily determine the small town where I live, what my interests are, what industry I work in, that I have written books, what make of car I drive, what language I speak, what music I like, what phone I use, whether my preference is PC or Mac, what university I went to, and a myriad of other “anonymous” data points.Google itself also uses third-party cookies in various products – though specifically not Google Analytics. The fear is that companies such as Google (also Apple, Microsoft, Yahoo, Firefox, Amazon etc) that have so much anonymous information about anonymous user’s that they can triangulate data points to identify an individual.

As you have probably concluded, it would not require a rocket scientist to be able pinpoint exactly who I am and identify me. This is why the EU law makers are trying to nail this down – and quite rightly. We need politicians and policy makers to protect user privacy in this way. The problem is that data triangulation often gets confused with any and all types of benign tracking that take place – such as that used by Google Analytics.

© Brian Clifton for Measuring Success, 2013. | Permalink | 7 comments | Add to del.icio.us
Post tags: cookie, cookie law, eu, privacy

What is this new feature…?

As described in their official blog post, this feature is aimed at organisations that transact in multiple currencies e.g. USD, GBP, Euros, SEK etc. Google’s recommendation is to create a separate copy profile were your transaction amounts are converted into a single currency – say all USD. This is achieved by pulling a currency conversion rate from Google Billing and automatically applying this to your transaction data in your profile. The conversion rate is the daily exchange rate of the day before the Google Analytics hit date.

This sounds like a good thing, right? I get all my transactions converted into a base currency so I can see a global overview of revenue performance…

Why this is flawed

Superficially this approach sounds like a good idea, but it’s flawed. Let me explain my thinking by posing the following question:

How does collecting data in multiple currencies, adjusted according to daily exchange rates, help me benchmark/optimise my website’s content, or its marketing efforts?

For example, if my “standardised” revenue starts to go down, how will I know if this is due to a poor conversion process (page content, load times, usability issues etc.), or a defective campaign? It could just be a fluctuation in the many exchange rates. As a case in point, the GBP recently dropped 15% against your Euro in one day. See also http://www.bbc.co.uk/news/business-16424802.

What’s the alternative…?

I use the following, trusted approach for multi-currency websites:

• Create a separate rollup account for your combined data.This includes combining all data – pageviews, events, and transactional data. I detail the rollup technique in an older post (also see chapters 6 and 9 in the book for a more up-to-date description). It’s pretty straight forward and essentially is just adding a second carbon-copy instance of the Google Analytics tracking code to your pages, but with a different account number.

• For rollup transaction data, use a FIXED currency exchange rate.
  By fixed, I mean set once and forget. This is the most important part – do NOT adjust for currency fluctuations. Therefore, if your business reports globally in Euros, convert your USD transactions into Euros when the Google Analytics data is collected. For example, Euro_amount = USD_amount x 0.75, could be your fixed conversion rate.

Using this method, there is no complicated “backing-out” of exchange rate fluctuations to do (which is a very complicated task in the first place) in order to know whether a change in the data is a good thing e.g. your marketing is working, or not.

In this way, the performance of your website(s) is what you measure
- not the performance of the currency markets.

Why this should bother you

Regardless of the audience, a message I always attempt to get across is the importance of using data for “actionable insights”. That means if you cannot take action on the data, or gain insights from it (i.e. a better understanding of your website), then that data is just noise.

Any feature that adds to the noise (and causes confusion!) needs to be avoided. Unfortunately that’s what this new feature does by default.

There is so much data noise available from even the smallest and simplest of websites that often people feel overwhelmed. It results in measurement paralysis and so the organisation makes no serious investment in web analytics because of this. This is a tragedy in my view and it still holds back the web measurement industry.

BTW, be aware of the other Google flaw I comment on – the issue of “not provided” keywords. Whilst both of these are important, lets keep this in context. This post is not about Google bashing, its about ensuring those of us using Google Analytics are able to make informed decisions about our data… After all, we do all love actionable and insightful data

If you feel Google multi-currency support feature really is advantageous to your organisation, please let me know your thoughts via a comment.

© Brian Clifton for Measuring Success, 2013. | Permalink | 3 comments | Add to del.icio.us
Post tags: flaw, flawed, multi-currency, multicurrency, not provided, roll up, rollup

Essentially, the change was that if a visitor is logged in to their Google account when performing a search (for example, logged into their GMail or any other Google service), Google takes this as a signal to request privacy – and therefore encrypts the session via SSL. The result is that when a visitor clicks on an organic result, no referral detail is passed to the receiving website i.e. the keyword information is lost – see technical note at the end of this post for more information.

Note: This effects ALL web analytics tracking tools – including Google Analytics.

The following chart plots the scale and growth of the “not provided” issue for 7 English language websites I work with. The total organic traffic analysed over the period is ~78 million visits. The legend is in order to match the graphed  lines.

Figure 1 - Percent of organic visits with “not provided” set

Figure 1 indicates there are 3 different zones – corresponding to 3 different types of users. I label these as A, B and C on Figure 2

Figure 2 - same as above chart with three zones identified

• Zone A:
  I consider zone A an outlier. It represents this website, which we know is targeted to Google users. Given that, I am surprised it is not 100%. That said, it does not represent other non-Google specific websites (yet!).
• Zone B:
  I consider these to be tech-savvy users that are most probably logged into a Google service, or search directly via the browser’s omnibox.
• Zone C:
  This is the inverse of Zone B i.e. non-tech-savvy users.

Figure 3 – adding a “control” to the data

In Figure 3, I add data from two highly respected universities – represented by the gold lines for US and UK (in order of the last data point). For these universities, I am assuming visitors to these websites are a broad mix of tech savvy users that are most probably logged into a Google service (or search directly via the browser’s omnibox), and non-tech-savvy users (the inverse group).

The golden lines fit nicely between zones B and C and hence I use this as my benchmark. Figure 3 shows that above the golden line(s), the audience can be described as more tech-savvy, while below it they are less tech savvy. At present (March 2013), the “not provided” benchmarks are 40% and 35% respectively for US focused and UK focused websites respectively.

The Latest Changes To Affect “not provided”

Clearly loosing the keywords visitors use to find your website is a big loss to any digital marketer. And although +Matt Cutts originally commented the change would only effect a small proportion of your traffic, clearly this was going to increase over time. After all, logging into a Google service is exactly what Google would like all its users to do…

Now the browsers themselves are also impacting “not provided”. In July 2012 Firefox announced a switch to SSL for all Google searches, and Safari follwed in September 2012. On January 18th, Google announced the same change. That is, even if the visitor is not logged into a Google service, the Chrome omnibox uses SSL for the visitor’s session. So another step to seeing organic keyword detail all but disappear form your traffic reports. SEO is definitely getting harder…!

Figure 4 - Proportion of global web traffic by browser type

Technical Side Note

Although the SSL protocol strips ALL referrer detail from http headers, the referring domain can be retrieved by browsers that support meta referrer (currently only Chrome). That means, https://www.google.com/ will show as the referrer for example, though the query parameter containing the search term is not available. To cater for browsers that do not support meta referrer (currently IE, Safari and Firefox), Google redirects the visitor through a http referrer with the visit search terms removed (q parameter is null).

Privacy Side Note:

What I find odd (and disconcerting) is Google’s approach to AdWords traffic. That is, Adwords visits to your website are not affected – you still receive the keyword detail. It is strange to me that Google considers privacy important for organic searches, but not for paid searches.

Thanks to…

+Per Pettersson for spotting the recent Chrome change and +David Vallejo for his excellent technical support. Prof. Ingemar Cox and Dan Jackson of University College London for sharing their GA data.

© Brian Clifton for Measuring Success, 2013. | Permalink | 26 comments | Add to del.icio.us
Post tags: flaw, flawed, multi-currency, multicurrency, not provided, privacy, seo

Panel = 3 TMS users + 1 vendor. Tag Management Systems – in my opinion the next big thing for web measurement setup (i.e. 2013), but also for the web development industry in general – deploy the necessary tracking code and program widgets across your site(s). Essentially, they are the CMS for your page tags. However, Jim made a very pertinent point on how these new vendors are positioning themselves:

“Why would IT willingly remove themselves from control of such a valuable part
of website functionality? i.e. pass over to a marketer” - Jim Sterne

I actually think this is bad positioning by the TMS industry – Marketing should not be the drivers, or the target audience, for a Tag Management System. After all, we are still talking about the deployment of code, and that is very much an IT role. Also, the ability to ring fence page tags and simplify the edit process is very attractive to an IT manager.

Reasons given for using a TMS:

• Speed of code deployment
• Flexibility to switch in or out vendors tags i.e. adding new widgets, change from WebTrends to Omniture etc…
• Interestingly, the panel could not quantify the cost:benefit of using a TMS.
  Typical comments – “its difficult to quantify”, “it makes like much easier but I can’t quantify the ROI of doing it”. In my view, it has to become a no-brainer if adoption is to grow. That will hopefully happen now that Google has joined this industry with their free Google Tag Manager (see eConsultancy’s post on this)

Privacy Discussion

Q&A with Vicky Brock (questioning) and Dave Evans (answering) form the ICO – the UK data privacy authority.

• ICO – received var 500 complaints so far (after approx. 18 months) about website privacy in the UK. Thats a lot for them, but in the grand scheme of things that sounds quite low to me.
• No non-compliance fines issued by the ICO to date. Its considered the last resort. ICO would much rather help you become compliant than punish you – assuming any non-compliance is not a horrendous abuse of privacy.
• There is a 30 page guidelines document from the ICO on best practice (BC: this now pretty good!)
• The FUTURE (from the ICO): There is a European legislative “move” towards explicit consent i.e. no longer implied consent allowed. Time-scale = next few years. Nothing is fast at the European level!

Other observations from eMetrics (in no particular order)

• Virgin Media make £4 billion revenue per year, 20% online. That’s a big number…! (Richard Phillips)
• Request from Andre Morys speaking about Conversion Optimisation – Understand what customers want form their web experience.
  • For example, please, please, please change how you display a contact form when a visitor uses a mobile device i.e. make it much simpler & shorter.
  • Don’t analyse in a silo.

My comment: Working in a silo is one of the biggest challenges for the web measurement industry. The issue with working across teams is that you need a senior level “sponsor” to co-ordinate and drive this throughout an organisation. A present, such buy-in (beyond, “yeah just go and do it”) is very rare. Related post from me: Why Web Analytics Users Are Falling Behind The Industry.

• The number people complaining about the cost of Omniture products was up this year (it goes up every year!)
• I am aware of the Queen of Search, but now I have discovered the Queen of Segmentation. Great insights from an expert practitioner.
• “Old school” – defined as Business Intelligence and Analytical CRM (Richard Phillips)
• Word of the day = Fiduciary – a new one for me. Thanks to Vicky Brock.

OK, I have run out of bullets now…

Also read my colleague’s @JohnWedderburn live notes form the day http://bit.ly/U0s4wl

© Brian Clifton for Measuring Success, 2012. | Permalink | One comment | Add to del.icio.us
Post tags: e-metrics, emetrics