Business Performance Conversion Testing Customer Experience Product Development UX

Deceiving your customers: dark patterns or bad design?

(This post originally appeared on the carwow Product & Engineering blog on Medium)

You may find this hard to believe, but not all people behind your favourite websites have your best interests in mind. Some, you might even say, are out to squeeze every last penny from you before you leave.

Ads covering every spare pixel are tiring and overlays can be downright condescending, but at least they’re being upfront about their intentions. The sneaky ones, the ones you can’t easily see, the ones built into the design – the ‘dark patterns’ – perhaps they’re the worst of all.

What are Dark Patterns?

Sounds like some kind of conspiracy theory from the dimly lit corners of the web, but ‘dark patterns’ is the term coined for practices on the web that deceive users into doing something unintended. You expect one thing, you get another., a site dedicated to educating people on the practices and naming and shaming offenders, describes it as:

“Normally when you think of “bad design”, you think of the creator as being sloppy or lazy but with no ill intent. This type of bad design is known as a “UI anti-pattern”. Dark Patterns are different?—?they are not mistakes, they are carefully crafted with a solid understanding of human psychology, and they do not have the user’s interests in mind.”

There are some great examples of this in the library on that site, but I want to share a possible dark pattern I personally encountered last week.

Colour me bad

Now on the grand scale of things, this one isn’t in the ‘Super-Evil’ category, more in the ‘Oh, that’s cheeky’ tier. And I feel a little bad naming them, as I liked the product and it was very helpful to me on a recent research project, but…

Informizely, formerly known as Insitez, offer a number of online survey tools?—?I’d used their very useful exit survey product to help understand why users were leaving a particular page on our website. It served it’s purpose, but I didn’t need it any more and didn’t want to keep on paying the monthly subscription fee.

So I logged into my account and opted to unsubscribe. This was the message I was presented with:

colour based dark patterns

It seemed so simple, but I hesitated. My brain went ___________. What had stopped me? It was only a fraction of a second and as soon as I realised I had to chuckle.

The colours were the wrong way round. Accepted norms in colour psychology say ‘red means stop, green means go’. If I want to proceed I click the green thing. Don’t I? Er, no.

Were they trying to trick me into keeping my subscription or could it just be ‘bad design’? The latter is possible, I suppose. A quick straw poll of a handful of designers in the office was met with laughter.


I’m guessing they’re not convinced. I’m just not sure what the thought process could have been behind the innocent selection of those colours and labels.

We’ll never know. I cancelled anyway. Using the red button. (that was the right one, wasn’t it?)

Lesson Learned

What can we take away from this? To borrow from Google… ‘Don’t be evil’. If you’re charged with growing a business, employing dark patterns might deliver you short term gains, but it leaves a sour taste in the mouth of your duped (/nearly duped) customers, serving only to build long term distrust in your brand.

**Bonus Dark Pattern**

In writing this I realised I encountered another potential dark pattern recently, buying trainers on the Sports Direct website. has a category for this type?—?the rather descriptive ‘Sneaky into basket’.

I selected a pair of Astroturf trainers, went to the checkout to pay, only to do a double take when I realised a magazine had automatically been added to my basket. It was only £1, but I certainly had no interest in it and was peeved that it had been added.

I wonder what percentage of customers complete their purchase completely unaware of this surreptitious addition?

shopping basket dark patterns

Business Performance Conversion Testing Customer Experience Product Development

Don’t let the HiPPO drive the bus

This article first appeared on the carwow blog on Medium.

Gut feel and opinion go a long way, but it can also take you down the wrong path if you don’t listen to your audience. At carwow, we’re strongly opinionated but it doesn’t rule our decision-making. We use data and user feedback to inform our opinions and decisions about product development and we’re better off for it.

Where do the Hippos come in?

‘HiPPO’ is a term I first heard 5 years ago. It an acronym for ‘Highest PaidPerson’s Opinion’ and was used to describe how many companies make decisions about their product development and marketing. When faced with a decision, the choice was made based on the opinion of a senior figure, with an absence of data or a conscious decision to ignore it.

My view of a HiPPO is less about the seniority and pay grade of an individual and more about the mindset. It’s very easy to believe we know best, or to assume “I could easily be our target customer” but it can be a dangerous habit. I’ve been working on websites since the last millennium (1999?—?yes, I’m that old!) and I’ve lost track of the number of conversion optimisation tests where I’ve guessed the wrong winner. You just don’t know. You need more information.

So if guessing is not an option…?

If you can accept that you don’t know all the solutions to your business problems, then the answer is very simple. You listen.

There are cryptic clues, hidden messages, explicit complaints, enthusiastic endorsements all over your business and from a wide variety of sources. Sometimes you just have to shut up and listen for them.

It’s a trait that’s strong in those who contribute to developing the product at carwow.

So how do we listen?

The information you seek comes in many forms and so you need to listen in different ways. This list isn’t extensive but does highlight a few of our methods.


To see what is happening we use a variety of data reporting sources?—?primarily MixPanel, Google Analytics and our own internal customer data. We can see who is using our website, how they interact and convert, and identify bottlenecks that are hindering growth.

On a more granular level, we can review pages using clickmaps, heatmaps and (one of my favourites) video recordings of users interacting with specific functionality on the site.

Feedback methods

Analytics is great at the ‘what’ part, but pretty lacking when it comes to understanding the ‘why’Why are users bouncing from your homepage? Why are they dropping out on your signup form? Why are they unsubscribing from your beautifully crafted emails?

Listen. They’re telling you why.

User testing sessions

These are fun, try it. With user testing, you ask a participant to carry out tasks on your website, whilst you sit next to them and observe. They think out loud, telling you what they’re doing, why they’re doing it and how it makes them feel. All the while, you’re screaming inside your head:

“Click the big green button! It’s right there in front of you!!”

It’s an enlightening experience that really underlines the (potentially incorrect) assumptions you’ve made about how your product should function.

Don’t pass over this method thinking you don’t have the capability or experience to run these sessions. Instead, pick up a copy of Steve Krug’s book ‘Rocket Surgery Made Easy’. It’s a quick and immensely valuable read that will have you testing within the week.

Customer visits

Unlike user testing sessions which can be done in your office or in a testing lab, you can also learn a great deal by observing your customer in their own environment. carwow is a two-sided marketplace, matching car buyers with dealers, so we often visit the dealerships to see how our business customers incorporate our service into their working lives. You quickly spot the holes you can’t see from your desk.

Customer feedback loops

One of my favourite elements about the structure of carwow is that the Customer Service team are part of Product Development. If our goal is to build a product that enables the best car buying experience you’ll ever have, anyone needing to interact with Customer Service is evidence that we must continue to find ways to make it better. And if you’ve ever looked at the information handled by a Customer Service team then you’ll know it’s an embarrassment of riches for anyone looking to improve their service.

First places to look? Dive into your helpdesk enquiries, chat session data and your phone call transcripts. You’ll find problems, solutions, ideas and emotions.

Exit Surveys

When you look at analytics data, the most depressing moment is when a visitor leaves, task incomplete. Why? What put them off?

We’ll never know, is the natural assumption, they’ve gone. Well, true, but we can always ask them as they head out the door. Add an exit survey to your problematic page?—?a little slide-in single question survey box that appears as the user’s mouse moves perilously close to the browser menu bars?—?and ask them what’s wrong.

We did this on a signup form on carwow, asking:

What made you decide not to sign up today?


The results were illuminating, giving us some quick-fix issues to address and food for thought for some conversion tests to run. Handily, they’re also a great tool for capturing the voice of the customer, so you can use the same terminology they use in your web and email copy and not the usual jargon-laden marketing babble.

Insights as far as the eye can see. Now what?

I think you’ll agree, if you’ve looked at your analytics, spoken with customers and observed their actions, then you’re in a better, much more informed place to make decisions.

At carwow, all this information feeds into our design and development process. Our product managers, UX designers and developers assimilate this knowledge into our designs and code.

The proof of the pudding is in the eating, as they say, so we A/B test major changes with our audience, ensuring that the choices we made based on their feedback were the correct ones.

And if not, we go back to the information, we listen some more and we iterate.

Business Performance Conversion Testing Customer Experience

Is Google’s No Captcha reCaptcha A Conversion Killer?

Spam is a pain in the arse. That’s pretty much something we all agree on. Some of it’s destructive; some malicious; most of it pointless; clogging up email; bloating CRM systems and messing up tracking and reporting.

reCAPTCHA, the most common counter measure, has long been a frustrating user experience, resulting in many an abandoned attempt to submit a web form. So when I heard about Google’s re-imagining of the reCAPTCHA I had to take a look.

So what is Google’s No CAPTCHA reCAPTCHA?

In announcing the release on their online security blog, Google claim that their new API, which ‘radically simplifies the reCAPTCHA experience‘, will enable users ‘to securely and easily verify they’re human without actually having to solve a CAPTCHA. Instead, with just a single click, they’ll confirm they’re not a robot‘.

Sounds great. Check out their video, it’s worth it for context:

Sounds a little geeky, but I was genuinely excited to give this a try.

So does Google’s No CAPTCHA reCAPTCHA reduce spam?

To test it, we implemented the reCAPTCHA on a lead generation form on our online accounting website.

Google reCAPTCHA on lead generation form

And the impact? Did it reduce spam?

Yes. It did, but it also had a rather painful unintended consequence – it killed genuine conversions.

In the interest of fairness, it’s worth saying this isn’t the most robust of evaluations – ideally tests should run longer, with more data. It ran for a full 24 hour period before we removed it. Given the numbers, I wasn’t prepared to keep it running.

So what happened?

If we compare the 24 hours with the reCAPTCHA versus the 24 hours prior without the reCAPTCHA, I can confirm it had the desired effect on the spam. Gone. Happy days.

Regarding conversions, lets take a look at some observations from studying the page and form data:

  • The pageviews for the two days were very similar
  • The percentages of unique users clicking the form submit button (‘Get started’) were practically identical over the two days

This suggests that the presence of the reCAPTCHA wasn’t off-putting to visitors. Those who wanted to sign up, would still try and do so regardless.

So far, so good.

The problem comes when you dig further into the conversion data. The number of successful form submissions FELL by 73% from one day to the next.

If we look at the number of unique users clicking on the form submit button and compare against the total number of clicks on the button we find our problem. Once the Google reCAPTCHA was added to the page the number of times the button was clicked DOUBLED – from an average of 1.7 times to 3.4 times.

Why? I can’t be sure, but I have a theory.

3 reasons why No CAPTCHA reCAPTCHA kills conversions

1. The appearance isn’t customisable

As far as I can tell, the appearance of the reCAPTCHA is not customisable. It can only look like this:


As such, I don’t think it looks like part of the form. It looks more like a banner.

2. The checkbox is unconventional

I’m not convinced the ‘check box’ looks like a typical check box. It just looks like a square and as such a user may not be aware that they have to click it.

3. The copy is ambiguous

I have a concern about the ‘I’m not a robot‘ text. Does the average person know what a ‘robot’ actually is in this context? Considering you’re reading this, I’m assuming you’re a web marketer or pretty tech-savvy. You know what a robot is, but are you an ‘average’ user?

The text also doesn’t have any instruction. Nothing tells you to check the box, so you have to work it out. In the words of Steve Krug, don’t make me think.

As such, my suspicion is that the ‘banner’ is ignored. The user believes they have completed the form and clicks the submit button. When nothing happens, the user continues to click the submit button thinking something is wrong with the website. This causes the number of form button clicks to double.

After a few clicks with nothing happening, they abandon the form and leave the site. Reduced spam, reduced customers.

Good effort, must try harder

I think Google’s intentions here are good. Everyone hates those old reCAPTCHAs. I’d happily never have to squint and cock my head to try and figure out another blurry photo of a house number just to access an online account. The new one does provide a better experience. I love the idea that they’re looking at the user’s entire engagement with the process – especially the mapping of a human’s jerky mouse movements that a robot struggles to recreate – but I think the execution needs more thought.

Personally, I’d rather any solution completely takes the onus off the user and is dealt with in the background by the technology. Less friction is not as good as no friction.

So, to quote their headline, ‘Tough on bots, easy on humans‘… yeah not so much Google. There’s a little more work to be done yet.

Business Performance Conversion Testing Customer Experience Innovation Product Development Search Engine Marketing

How to growth hack your business and achieve sustainable results

The first rule of Growth Hacking? Don’t talk about growth hacking.

That was my paraphrased tweet quoting one of the speakers at the Growth Hacking Conference in London recently. It seems I’m not the only one who isn’t a fan of the name but feels the principles behind it are sound, if a little misunderstood.

Many people (maybe just us Brits?) roll their eyes when they hear the phrase ‘growth hacking’, thinking it’s a cheesy term applied to the processes of inflating numbers exceptionally quickly with a goal of reaching a ridiculous valuation figure that gets Mark Zuckerberg jingling his pocket change.

So it was refreshing to hear so many of the speakers at the conference talking about sustainable growth – yes, build a product that scales, but also one that provides value to its audience, not just for 3 days, 3 weeks weeks or 3 months, but on an ongoing basis. A product that would be missed if it were no longer there.

I scribbled a lot of notes throughout that day, so let me share a few of them here, along with the speaker slidedecks. If you want more, head over to the collaborative Google doc that was edited live throughout the event.

Find Product / Market Fit

The Father of Growthhacking (or at least the phrase coiner), Sean Ellis (@SeanEllis) hosted the event and shared his thoughts on the drivers of growth:

  1. Right product in the right market and
  2. Successful execution of ideas

Product / Market fit was a theme throughout the day. Ellis posits that without it, sustainable growth just isn’t possible. In fact, it’s critical for strong organic growth (Growthhackers are particularly keen on organic growth. Most argue the paid stuff should come once you have traction. That’s not to say you can’t scale a paid channel to achieve growth. You just need deeper pockets).

Ellis has a useful way of determining if a business has a strong product / market fit. He surveys the customer base and asks how many would be ‘very disappointed’ if they couldn’t use the product anymore. If you can hit 40%, then you have a strong product / market fit. Then you need to stack the odds in your favour by understanding what those people are getting from your product so you can reproduce it / scale it with others.

Hustle to find scaleable growth

Zack Onisko (@zack415), from Creative Market, talked of trying out creative, out-of-the-box ideas to get to your goal in the shortest possible time. He referred to it as the Hustle stage – trying out a variety of channels, potentially non-scaleable, to try and find one that could be unlocked and scale.

Gaining Early Traction by Doing Things that Don’t Scale (Hustling) from Zack Onisko

He gave 20 ways to hustle, of which my favourites were:

  • Speaking with customers to find out how to improve the product
  • Go above and beyond with customers – sending them surprise gifts, thank-you messages, building brand loyalty
  • Give something away for signing up – asking for info is high friction, so do an exchange. They give you their email address, you give them something of value to them (i.e. a book, free service, money)
  • Forge partnerships with other likeminded businesses – help each other
  • ‘Eat your own dog food’ – make sure you use the product yourself. Understand it, know its weaknesses and improve them

Build a positive Net Promoter Score

NPS Driven Growth or How to grow your startup w from Nilan Peiris

Net Promoter Score (NPS), was discussed several times, particularly by Nilan Peiris (@nilanp) of TransferWise, as an effective tool for measuring how your audience views your brand or product. NPS asks a single question:

How likely is it that you would recommend us to a friend or colleague?

Customers answer on a 0-10 scale and labelled Detractors (0-6), Passives (7-8) or Promoters (9-10). To calculate your company’s NPS, take the percentage of customers who are Promoters and subtract the percentage who are Detractors.

Clearly you want as many Promoters and as few Detractors as possible. How do you do that? Build a ‘must-have’ product, provide great service and exceed expectations.

Be aware though that a large number of Passives isn’t ideal either as those people are indifferent to your product, meaning they’d have less resistance switching if an attractive alternative product was available.

Study your successes

It’s very easy to get caught up trying to figure out how you can persuade the disinterested to buy your product. An alternative approach was suggested by Dr Karl Blanks from Conversion Rate Experts. He suggests it’s more productive to focus on those that did buy. He recommends asking customers immediately after they have purchased this question:

What was the one thing that nearly made you not buy from us today?

Whatever their one thing was, it wasn’t enough to dissuade them, but it could be putting off many others who are dithering over their purchases also. If you can identify those issues and fix them, you’ll win those ‘almost-in-reach’ customers. That has got to be easier than trying to convert someone not interested in buying.

A bonus tip from Blanks: reverse engineer success by creating a custom segment in Google Analytics for those that bought from you. Then look at their behaviour and the content they looked at and try and figure out why those were successful.

Find your customer’s voice

SEO Tactics to Love vs. Leave from Rand Fishkin

Moz‘s Rand Fishkin (@randfishkin) and his legendary coiffed hair adorned the stage to speak about organic SEO growth. His slides were full of great nuggets but if I’m only to pick one, I found his thoughts on using the customer’s language thought provoking. He discussed ‘intent keywords’ and how to build a list of them to target with content. Keyword suggestion tools have their limits, so he recommended these three activities to generate them:

  1. Look at your competitors activity – see what they rank for
  2. Talk to customers directly (and your sales people) – what words are used in the conversations
  3. Look for conversations online in forums and places like Reddit – what is the language used?

It’s surprising how often businesses make assumptions about the words to target for SEO and content. The gap between assumption and reality is often a chasm. Locate the words used by the customer, not you.

Other Awesome Speaker Decks

There were so many great speakers at the Growth Hacking Conference, it would take quite a lengthy post just to summarise the main takeaways. So instead I’ve rounded up the remaining slidedecks I could find and included them here.

Hands-on Hacks – A Flipbook of Inspirational Growth Wins from Marie Steinthaler

The 3 Ugly Ducklings of Growth. from Geckoboard

 A Growth Checklist

As the day closed, I tried to summarise the main takeaways in a list to take back to the office. So in my newly educated opinion, If you want to build sustainable growth for your business I would suggest:

  • Build an awesome product (find Product/Market match)
  • Develop a strong Customer Value Proposition (i.e. why you’re relevent, what you offer, why you’re better)
  • Provide great customer service
  • Build in feedback loops to your product (and act on it)
  • Find likeminded partners to promote
  • Ensure data / measurement are baked in to your product and processes
  • Optimise the journeys (using conversion rate optimisation testing, i.e A/B or MVT)
  • Delight your customers (and measure with NPS)
  • Find smart people to advise you (and to ground you when you fall in love with your product and can’t see when you’re wrong).

So, all in all a great conference, one of the best I’ve attended. There aren’t too many events you can attend where you can discuss product, data, conversion rate optimisation, customer service, partnerships, entrepreneurship, with smart people who are equally happy to talk about their mistakes as well as their successes. If its on again next year, I’d recommend you attend. I’ll see you there.

Awesome photo courtesy of Dan Barker, ecommerce & online marketing expert and dab hand with a camera. Worth a follow on Twitter @danbarker

Business Performance Videos

Watch what happens to your web traffic when your TV ad airs

Have you ever wondered what impact a TV ad has on a website when it airs?

I don’t mean overall impact, such as an increase in brand awareness or revenue. I mean the IMMEDIATE impact.

As in, if my advert appears on TV right now, what happens on the website?

Let’s find out

The introduction of Realtime into Google Analytics was a wondrous thing for digital marketers. It provides an opportunity to get a feeling for what is happening right now on a website, without having to rely on day old data. It can be very exciting (and a little mesmerising).

When, my previous employer, ran a new TV campaign in January 2013 it presented a perfect opportunity to answer my question of ‘what happens next?

At 9.45 pm on Sunday 6th January, one of our adverts appeared in the commercial break of the popular drama, Mr Selfridge, on ITV1 (SouthEast & Midwest regions only). Using video capture software, I recorded the Realtime page of our Google Analytics account as the TV ad played and for the 4.5 minutes afterwards.

TV ad impact on web traffic

So what happened?

As the TV advert aired

TV impact on web traffic - at the start

When the advert started there were 1,574 people on the website at that very moment. 51% of these visitors were New and 49% were Returning (had used the site before).

Within seconds of the brand name being mentioned and the website URL appearing at the end of the advertisement, the number of people on the site started to rise.

1 minute after

TV impact on web traffic - after 1 minute

After a little over 1 minute the number really started to climb. I suspect that is how long it takes people to reach for their mobile devices or laptops and type in the URL or conduct a search. At the same time, the split had shifted to 55% New visitors versus 45% Returning visitors.

2 minutes after

TV impact on web traffic - after 2 minutes

About 2 minutes after the ad finished the number of active visitors had increased about 1/3 on the starting figure to 2,084. At that point the split had shifted to 57% New visitors versus 43% Returning visitors.

3 minutes after

TV impact on web traffic - after 3 minutes

The visitor numbers peaked at about 3 minutes after the ad ended, hitting 2,143 active visitors. The visitor split was 58% New visitors versus 42% Returning visitors. By this point the programme was back on and numbers started to settle back down again.


It’s worth mentioning that this was just one ad slot on a Sunday night early in the campaign and isn’t necessarily reflective of the campaign as a whole. However, it does provide some interesting observations:

  • TV advertising drives people to your website. A no brainer really, but nice to see it works.
  • The speed at which people started visiting the site supports the notion of increased adoption of ‘dual screening’ – people using their mobile devices (or laptops) whilst watching TV.
  • TV advertising attracts new users to your site (evidenced by the New/Returning split change), as well as bringing back previous users.

The campaign itself was a great success, breaking records across a range of metrics including visitors and job applications. It also massively helped shift awareness of the Jobsite brand. When we started TV advertising in 2008, awareness levels amongst jobseekers was only 31%, but rose to 68% by the end of the January 2013 campaign. It was equally as impressive with the recruiter audience, rising from 35% to 67%.

As a digital marketer, it’s rare to get an opportunity to do brand focused activity on such a large scale. It also takes some getting used to the ‘measurement’ systems used, such as TVRs, brand consideration, empathy and awareness,  so it’s great to be able to make use of our digital tools to make sense of the impact that offline activity can have.


I originally wrote this post for the corporate blog of Evenbase, the parent company of Jobsite. Following the closure of Evenbase, the article has been reproduced and updated here on my own blog – don’t want a good post go to waste 

Business Performance Conversion Testing Email Marketing

How to recover 52% of your abandoned shopping carts using email remarketing

When the subject of conversion optimisation comes up, most of the talk is about on-page optimisation – utilising A/B or multi-variant tests to determine which headlines, calls to action, images or page layout persuades your visitor to take the necessary steps to reach their goal (and yours).

It’s not the only way, mind you. There are ways to increase your conversions using channels outside of your website.

Let me paint you a picture.

Your company sells car insurance and I’m shopping around for a new policy. I dread this point every year and in fact it’s often the reason why I just renew my policy even though I could get a better deal elsewhere. But this year is different. I’m going to find the best deal and it just so happens I’ve found it on your website. 

I’m pretty pleased, that didn’t take long. So I start to fill out your online application form. The smile on my face slowly begins to fade. Within five minutes it’s a full on scowl. So many questions! You’re asking for dates for this, policy numbers for that and… oh my word, is this really worth it to save £40???

Thing is, I know that every insurance website will ask me this. If I want the saving, I’m just going to have to go dig out the old paperwork so I can answer the questions. But I don’t have time right now, I’ll do it tomorrow…

Or not, as so often is the case.

It’s not just insurance, it happens everywhere on the web where some kind of transaction is occurring. It’s typically referred to as an abandoned shopping cart, but it equally applies to non-ecommerce processes, such as lead generation forms, job applications and account registrations, for instance.

As optimisers, this pains us. It’s our job to convert as many visitors to customers as possible. We could just accept these as lost or we could figure out how to get them back.

Bringing them back

Remarketing, as the name suggests, is the activity of re-engaging with visitors who did not convert first time round. The most popular form of remarketing is via display advertising. A cookie is set when a user visits a site and if they do not convert then personalised display ads are targeted at them on their subsequent travels around the web. Major players of this technology include Struq, Criteo and Google.

That kind of remarketing can be very effective, but I want to talk to you about another type of remarketing that can do wonders for your conversion rate. Email remarketing.

Urgh, email marketing. Boring. With all the sexy new technology available, people would rather not talk about email. But you know what? It’s still one of the most effective marketing channels you can use.

Why? Because there’s infinitely less competition for attention within the private walls of their inbox than there is out there on the wide open web. Especially if the message is one to one, relevant and of value. For instance…

Hi Gary,

It looks like you weren’t able to complete your insurance application today, so we’ve saved the information you entered. When you’re ready, just click the link below to pick up where you left off and you’ll be insured in just a few minutes.

How I recovered 52% of abandoned shopping carts

Lets look at a real example. In a previous role I worked for, a leading job board in the UK. Advertising recruiters have a choice of buying job ad credits over the phone from a salesperson or they can do self-service via an ecommerce system, complete with shopping basket.

Tasked with increasing revenue, I analysed the performance data to identify optimisation opportunities. I spotted one in the basket abandonment data. These people were interested enough to place a product in the shopping basket and start to fill out their billing details… but for some reason just couldn’t bring themselves to hit the ‘Buy’ button. Whilst there were plenty of on-page optimisations to do, I felt these people were so very close to a conversion that we had to try something a little different.

So what did we do?

We decided to utilise the Conversion Manager product from SeeWhy, which provides real time personalised messages to visitors that abandon shopping carts. With the software integrated with our system we were able to send triggered emails based on the behaviour of the visitor and the information they provided on the site.

Each visitor could receive up to 3 emails depending on their actions, or inaction, all written in a non-sales tone and featuring a link to their shopping basket, containing their partially completed data.

The result? 52% of those visitors who received a remarketing email from us returned to the shopping basket and completed their purchase.

So, a little more about the emails…

The first email needs to be sent immediately. There are 2 reasons – 1) not all abandonments are deliberate, sometimes people get kicked out of systems or lose internet connectivity and 2) the visitor is in a buying frame of mind if they were in your checkout, so now is a great time to reinforce your message and offering.

If the customer does not purchase (or register, or download, etc.) within 23 hours then send the 2nd email. People are creatures of habit and your visitors are likely to be back online the same time next day (i.e. lunch break). Why 23 hours and not 24? Well, that allows for any delay in the ISP delivering the email and for it to be ready waiting for them when they next login in.

If they’ve still not purchased after 6 days and 23 hours send them the 3rd and final email. Any more than this and it starts getting spammy. You might want to trial a discount voucher code in this final email (we did), as one last attempt to get them to complete their order.

Be careful of offering the discount any earlier than this. If your promotion occurs in the first email, you’ll train your visitors to abandon the shopping cart immediately, just to get the discount email. It’s a very patient customer who will wait a week to get their third email just to game the system.

The tone of the emails is important. A friendly customer service tone is what you’re after. An overly aggressive sales message will come across as spam – and remember, some of the abandonments may be because of genuine difficulties completing the process.

Make sure you capture the contents of a field as the data is entered. This makes it possible to return the user to the point they left off when they click the link in the email. This is a big win for the visitor – tell them in the email copy that they can pick right up where they left off. This’ll increase the likelihood of them finishing the transaction.

One of the advantages of using a product like SeeWhy is that it synchronises with your purchase data in realtime. This ensures that you don’t send out a remarketing email if the visitor has recently returned to the site and made a purchase. Failure to stop the scheduled email would result in a confusing experience for the customer – and a frustrating one if you’ve sent them a discount code moments after they purchased the product at full price.

So how did each email perform?

EmailTimingShared of Recovered Baskets
Email 1Immediate78%
Email 223 hours later16%
Email 36 days 23 hours later6%

As you can see, the vast majority of recovered shopping carts came from the first email. If need be, you could just implement a single email if you have limited resources and you’ll still likely experience an uplift. However, you may feel that it’s worth the additional work to recover the additional 22% from emails 2 and 3. It’s also worth noting that you can experiment with the timing of the emails. The ideal phasing of the three emails could be different for your product and audience.

So, if you’re looking to optimise your shopping basket (or your lead generation form, your sign-up process, or download page), think about all the channels at your disposal. To encourage visitors through the buying funnel, you could A/B or multi-variant test the content of your web pages. To pull browsing visitors who did not convert back to the site try remarketing via display adverting. But to get those people who were so close to buying that they almost completed your shopping basket process, give email remarketing a go. Guessing a 52% recovery would look pretty nice on your stats sheet?

Business Performance Conversion Testing

What a 3 year old can teach you about conversion rate optimisation

We do like to make things complicated.

We see something simple and feel like its missing something. It’s been a while since we launched this, we should probably do an upgrade; customers expect new features; some bells and whistles would be nice.

Tell me, what couldn’t be improved with some extra bells and whistles??

Turns out an awful lot. If you’re in the business of driving action online, of increasing revenue or sign ups or downloads, then you should care a lot about providing a simple path to conversion.

Done well, changes to overly complicated pages and processes will help your conversions improve. Mess them up and you’ll find the arrow pointing in the wrong direction.

So how do you ‘de-bell’ and ‘anti-whistle’ your cluttered webpages?

It’s simple. Watch a small child playing with a jigsaw puzzle.

Er, say what?

Yep, close your laptop, leave your desk and watch how a child plays.

I can’t claim to know how they think. I was one once, but sadly can’t recall anything of the thought process of my 3 year old self. What I can see, however, is that their actions are not encumbered by complex thought, scenario planning, or indecision, yet their play is incredibly efficient.

Sounds like a recipe for conversion rate optimization to me.

Let’s take a look at four jigsaw puzzle takeaways:

Appearance counts

Children can be fickle creatures. With plenty of exciting toys to choose from, it’s pretty hard for the humble jigsaw puzzle to stand out from the crowd. The child will scan the room and make instant decisions on objects. So what makes the child pick a puzzle? What makes it stand out?

The picture on the box.

The child is drawn to the image – the bright colors, the composition and of course, the appealing familiar characters. A dull, featureless image just won’t cut it.

Takeaway: Know your audience and make sure your website design appeals to that group. Making your site credible is important, whatever your industry. That doesn’t have to mean formal or corporate, just appropriately designed for your audience.

When someone lands on your site, they’ll make a split second judgement based on appearance. Get it right and they’ll pick your site from amongst all the other sites in the Internet toy room.

Think about your use of colour, headline, imagery, layout and trust marks, but overall make it clear that it delivers on their need.

Speed & simplicity

Kids have a short attention span, so toys need to peak their interest immediately. If they’re too complex or confusing they’ll quickly lose interest. A child will do several simple puzzles, excited by the achievement of completing each of them, but will balk at a larger, more complex puzzle that requires greater concentration and effort.

Takeaway: Steve Krug wrote a great usability book called ‘Don’t Make Me Think’, the title of which should be a mantra for anyone tasked with building processes. If your webpage makes me stop and ask “What do I do next?” or “How do I buy this?” then something has gone horribly wrong.

Cut out the unnecessary stuff and just focus on letting the user complete the process as quickly as possible. Filling your customer database with a wealth of ‘enriching’ data is unlikely to be worth the trade-off of losing potential customers through poor process design.

Visual Clues

How does a three year old start a puzzle?

Scatter puzzle pieces on the floor and watch which piece they reach for first. Invariably it will be a face – Mickey Mouse’s big eyes or Bob the Builder’s bright yellow hat. They’ll latch on to these visual clues to give them a bearing. They’ll repeat this process until they can join pieces together.

Where they struggle is when they try to match pieces with indistinct imagery – green grass or blue sky – it just becomes a process of trial and error to see which pieces fit.

Takeaway: You can’t always afford to be subtle in your design. If you want your visitor to click a button to move into your purchase funnel, then make sure that button stands out from the rest of the page elements.

It’s all very well having a colour palette to adhere to in your Brand Guidelines, but if your button blends in it won’t do much for your conversion rate. Repeat after me: It’s okay to be different.

Don’t stop at colour, experiment with size, shape and location too.

Remove obstacles

A quick way to lose a child’s attention is to present them with an obstacle, such as a piece of the puzzle missing from the box. They won’t go looking for it. They’ll give up and go play with something else.

Takeaway: The same applies to your website. If your landing page or process has unnecessary obstacles – such as poor process flow, clumsy navigation, broken functionality or unnecessary actions then your users will give up early and go find your competitors. Remove the obstacles and your conversions will increase.

Two things to watch out for on your site:

  1. If you’re running an ecommerce site then trust is essential. If any part of your process breaks or behaves in an unconventional way then it will introduce doubt into your visitor’s mind. This could result in an abandoned shopping cart.
  2. Fix your form errors. As a user, it’s so frustrating to complete a form and then be presented with angry red error messages saying you’ve made a mistake. If you want my password to be 8 characters, containing at least 1 numerical character, then tell me upfront. If my telephone number has to have a space after the area code then tell me (actually, just figure it out in your code, you’re smart enough). And don’t you dare remove the info I’ve already typed in other fields when you’re busy presenting me with error messages. Test your form over and over again and try and break it. Don’t rely on your users having infinite patience.

The conversion payoff

Have you ever seen the look on a child’s face when they click that last piece of puzzle into place? They’re so proud of what they’ve done, they’ll probably want to do it again.

The same goes for your customers. If you make it easy for them to get the product or service they need, to make it a pleasurable experience not a painful one, then they’ll come back time and time again.

What other conversion lessons could you take from how a child views the world? I’d appreciate your thoughts.

Social Media Uncategorized

11 simple steps to scupper your social media strategy

This far in you’d think brands would’ve started to figure out this social media malarky. There’s plenty of best practice advice available, and dare I say it, common sense should help. But still, there are plenty of examples of how brands shoot themselves in the foot with their social media strategy.

There’s no cookie cutter approach that fits all brands (and that includes your personal brand), but there are certain things you should try to avoid. Lets look at a few here:

1. Randomness as a social media strategy

Some accounts don’t seem to have any purpose. Social media is part of your comms strategy, whether you’re using it for customer service, promotion, or education. Therefore, like your PR, customer emails, or advertising, there should be a plan as to what you’re trying to achieve and the messaging you’re trying to convey. Sharing random articles, posting sporadic updates or only sharing your latest blog updates is not a social media strategy and nor, is it an effective way of building an engaged community.

2. Hashtag hijacking

A surefire way of generating some negative attention is to jump on the popularity of a hashtag when there is absolutely no connection with your brand. This has been done to disastrous effect with tragic news stories, but the trend du jour seems to be hashtags around popular TV programmes. Most often it comes across as lame, desperate and accentuates the lack of conversation around your own brand.

3. Me, me, me

Like dating, it’s a turn-off if you just talk about yourself all the time. When you’re trying to grow your business its understandable that you use your social channels to share your blog content. But people have trouble connecting with brands at the best of times, so solely sharing your own content comes across as just self-promotional. Sharing informative, useful content written by you AND others, however, will make you a helpful, valuable resource.

4. Bitching with other brands

Tesco Mobile made a bit of a name for themselves by having an edgy approach to social media. They went against the grain and took on naysayers in a humorous way. Others brands started getting drawn into conversations and some fun conversations were had. Unfortunately, this appears to have emboldened some brands into being bitchy about competing brands. Sadly, it just makes the brand look petty. Know the line not to cross and play nicely with the other boys and girls.

5. Retweeting every nice comment made about your brand

Yes, it’s great when someone says something nice about your brand. And fair enough, sharing a little of those nice comments can be useful in reinforcing to others that your product / service / advice works. Just don’t over do it. Filling your audience’s social channels up with comment after comment quickly becomes annoying. And don’t even think about retweeting the retweets of your tweet!

6. Not replying to questions from people

With bigger brands, volume can be an issue – but not an excuse – when it comes to replying to comments. It may just be me, but when someone asks you a question, it’s rude to ignore them. You wouldn’t do it in a physical store, so why online? So if you’re going to be active in social then make sure that someone is there to answer the questions, as well as posting your content. Remember, it’s supposed to be a conversation and conversations need to be two-way.


Endlessly sharing famous or inspirational quotes is not a social media strategy. It just shouts “I’ve nothing interesting to say so I’m just going to pad it out with something ‘inspirational’ that someone else once said that has probably already been endlessly tweeted by other brands who don’t know what else to say”. Or something to that effect. If you want to inspire others, do it with your own original thoughts.

8. Flogging your wares

Imagine talking to a friend in the pub and someone tapping you on the shoulder and saying “I couldn’t help but overhear you mention the word ‘car’ and wondered if you’d like to buy some motor insurance?” You’d tell them where to go, right? So, why do some brands act the same way in social media. Monitoring software is a great tool, but be considerate when joining conversations uninvited. If you want to jump in, be helpful not sales-y. Be useful, foster goodwill and you’ll have a better chance of turning the prospect into a customer when they reach the appropriate point in their buying decision-making process.

9. Buying your friends

Trust me, if anyone promises to get you 5000 followers for $100 dollars, then just hit ‘report spam’. Spending money to buy popularity is a precarious social media strategy. Quality will be low and engagement non-existent. Far better to grow organically by being really interesting and really helpful. This is all beginning to sound like a guide to making friends in real life, isn’t it? Funny that.

10. Outsourcing your voice

This is a contentious one and plenty of big brands do this. There are agencies who provide a full social service, creating assets, sharing content and responding on your behalf. What they lack, however, is authenticity. Their voice is not your voice, it can be a close facsimile, but they’ll never have the passion that you’ll have for your own brand. They also will be detached from what is happening inside your business. A maintenance page going up quickly whilst a problem is fixed may not reach the agency in time, who are busy running sponsored campaigns driving traffic to your website. That disconnect can be costly.

11. Lies, Damn Lies and Statistics

Hopefully you’re monitoring the impact of your social activity. Preferably in a meaningful way. It’d would be great if you can show how it’s impacting your core business (i.e. generating leads / sales), but at the very least how it’s increasing engagement with your prospective customer base. Please, though, don’t just count followers or fans – if that is the only thing that’s important you might as well go buy those 5000 followers for $100. If they’re not interacting with you, does it matter how many have clicked the follow button?

Once you’re looking at engagement metrics consider the sentiment of those interactions. 4000 mentions from a tweet isn’t a good thing if 95% of them are negative. You need to balance volume with what is being said. Is your social media activity generating positive conversations? Are people sharing your content (because they like it, not because you made a mistake or did something shocking)? If not, then you need to review your approach.

As mentioned up top, there is no cookie cutter approach to social media. You can do whatever you like – and maybe some of these tactics will work for you – but just be considerate of your audience. The social part of ‘social media’ is a big clue. People are on there to mix and share with others. As brands we need to tread carefully on their turf. Timing, relevancy and value should be the watch words of your social media strategy. Better that than bull in a china shop.

Social Media

How to avoid the #SocialFAIL

If you were on Twitter this weekend, you’ll likely have seen a couple of brands receiving considerable attention for tweets posted by their official accounts. Whilst management at Tesco Mobile may well be grinning today, the mood will likely be a little more uncomfortable in the offices of The Sun newspaper.

The Sun Showbiz account, live tweeting during Saturday’s episode of X Factor, made an ill judged attempt at being funny and topical by making a joke at the expense of beleaguered contestant Tamera Foster, who is alleged to have stolen make up from Boots. However, drawing missing child Madeleine McCann in to be the punch line of the joke drew the ire of tweeters on Saturday night.


It took nearly an hour and a half, but the account issued an apology for its mistake.

sun apology

Conversely, Tesco Mobile received plenty of praise for its response to a tweet knocking the brand.

tesco mobile comeback

This tweet went viral – at the time of writing it had been retweeted 6,401 times. It was a very well judged response. The original tweet by the member of the public had humorous intent, so the Tesco Mobile staffer responded in kind, being more playful than most brands would dare.

Interestingly, it’s not uncommon for Tesco mobile. If you check their timeline you’ll perhaps be surprised by how interactive and irreverent the account is – it even playfully spars with its competition.

Your social brand

In looking at the accounts it’s clear that both brands have taken a brand-led approach to conversation rather than a corporate-led approach. They have a view of how they wish their brand to be perceived by consumers and have decided that engaging through the use of humour is a good way to reflect that. The trouble is everyone has a very different sense of humour and you get to a certain point with jokes where you lose part of the audience. The art is in understanding where that line is and knowing when not to cross it.

Many brands struggle to put personality into their social profiles. Most tend to want to play it safe for fear of something going wrong and there being a backlash. Others are braver, realising it’s that personality that sets them apart from their competition.

So what do you need to consider for your own social media accounts?

Know your plan

Do you know what you want to achieve with your social channels? Many consider it a customer service channel, others view it as a brand building route. Many use it for both. Like with all your activity, you need to be aware of your objectives before you start. This will shape how you behave.

Know your voice

If you’ve ever done any brand development work you’ll have gone through exercises that ask questions like “what does your brand look like?”, “what does it feel like?”“what does it sound like?” The last one is key in this instance. At the very least you’ll have a tone of voice document (and if not, write one), that details how you should talk to your audience in your comms and advertising. Your social media conversation is an extension of these activities. It should be the same voice.

You may look at some of the great stuff brands are doing in social media and want to emulate it for your own brand, but using edgy language and humour because others successfully do it isn’t good business sense if it doesn’t fit with your brand. Instead, know your voice and find creative ways to engage your audience.

Everyone on the same page

Once you know your voice, you need to make sure everyone on staff representing the brand in social media knows it too. It’s not something that should just live in the head of a marketer. If your Customer Service personnel use social media to assist customers then they need to be well versed and comfortable in the way that your brand will talk. Imagine how jarring it will be if multiple employees are talking to customers in different tones of voice. You don’t want robots, but you do want consistency.

Likewise, make sure that anyone operating a social media account for your brand is fully aware of how you want them to behave. If the brand calls for you to exhibit a ‘quirky sense of humour’ make sure your employees know exactly what that means. They need to be briefed, trained, maybe do some workshops to give them some practice. That way you don’t have a staff member manning the account alone over the weekend, posting something you might regret.

This is about empowerment, not restriction. I’m not a big fan of corporate social media policies. They tend to focus on what you can’t do. I prefer guidelines – advice on things to consider when posting; about using common sense. Empowering people to make good decisions comes down to clear objectives, good planning and plenty of support. It gives you your best chance of making sure everyone is prepared to respond appropriately in the moment.

Know what to do when it all goes wrong

Okay, so you really don’t want to get to this point, but I’d urge you to think about what happens if things do go wrong. Can you answer these questions now, let alone when your brand is under fire in the socialsphere?

  • Can you quickly get a summary of what has happened and who is involved?
  • What is the plan to tackle the crisis?
  • Who in your business needs to be informed?
  • Do you remove the content?
  • Do you issue an apology? And how?
  • How do you ensure it doesn’t happen again?

You don’t want to be making this up on the fly, because you don’t know who will be around to pick up the pieces and your response could determine how your brand is remembered.

Play nicely

As an aside, Tesco Mobile did very well out of this situation, but poor Jay Feliipe took a lot of flak, which was probably not their intention. So it was nice to see this little exchange later over the weekend, with Tesco Mobile reaching out to him to offer him something to make up for all the negative attention. A nice touch. I wonder what it was?

tesco gift

Update 21/10/13 22.20 pm:

Ha ha, Tesco Mobile certainly have a sense of humour. Jat Feliipe, taking the comeback with good humour has since tweeted out a photo of the gift he received from Tesco. For a young man with specific requirements in his dating life, he’ll probably appreciate a selection of male grooming products and a copy of The Rules of the Game by Neil Strauss. And I’d hazard a guess that’s a Tesco Mobile SIM card in the background!

Nicely done, Tesco Mobile, nicely done.

Thank you @tescomobile for the gift much love for you!!

— Felipe (@JayFeliipe) October 21, 2013

Product Development Uncategorized

Making an ass of your assumptions

When faced with a problem it’s very easy to go with the obvious answer. If we expected X to happen but instead we got Y, then we can only assume Z must be the reason.

If presented with a hard fact, we tend to fill in the blanks to make sense of why something didn’t happen the way we expected. Sometime we’re right. There is a reason why it’s the obvious answer. But like stereotypes, which can be based on repeated traits, assumptions can also be wide of the mark.

We need to think differently. We need to dig beneath the surface.

To do so, it’s important to understand the data, but we also need to put ourselves in our customers shoes and look at it from their perspective. Data can give you the what, but not the why.

Scrap it, start again

Let’s look at an example. My previous employer,, launched an iPhone job search app.

It was built in response to the considerable mobile traffic growth they were experiencing. It was a basic app, nothing fancy, but focused on the core need of the jobseeker – to search and apply for jobs.

It had several hundred thousand downloads, drove a lot of activity, but it only averaged 1.5 stars out of 5. This was pretty much in line with the other job hunting apps in the App Store.

The initial reaction to the score?

The app is not good – we need to either upgrade it or build a new one.

But that didn’t sit well with me. I didn’t think the score was a fair reflection of the app. Yes, I knew there were things that needed to be improved, and we were working on them in the background, but I felt the app was better than its rating.

Why was it so low?

Digging for treasure

It didn’t take too long to find the answer. Aside from the star rating, Apple allow customers to write a review of their app experience. When your app is getting positive reviews – 4 and 5 stars – then this is a great piece of social proof that is likely to encourage other shoppers to download the app. However, negative social proof can have the complete opposite effect and put prospective customers off.

Once you get over the initial sting of reading the reviews you realise how much of a gold mine of information it can be if you’re interested in improving the customer experience.

With a little reading it became apparent that most of the reviews were from people frustrated that they couldn’t find or get a job, and were projecting that frustration in their review of the app.

It’s an unfortunate fact within recruitment that most people are unsuccessful with a job application. At best, one person is going to be really happy (if they got the job), some might be disappointed but had a positive experience during the process (e.g. good recruiter service or getting short listed), but most are frustrated due to not even getting any feedback to their application let alone invited to interview.

The initial thought at this point is “well, we can’t do anything about that, that bit is out of our control”. Which is true, a job board plays little to no role in that part of the recruitment process.

So I guess we can’t affect the scores then? We’re stuck at 1.5 out of 5. Let’s move on.

Not so fast. Let’s look at this differently.

Finding out the why

We’ll start with user behaviour. When do users leave feedback on apps?

It’s actually not that easy to do. If a user is being proactive they could visit the App Store, find the app and leave a review. This does requires a bit of effort on behalf of the user, and I’d hazard a guess most people don’t have the time or inclination to do that.

So when else?

When you delete the app from your device.

And when are you likely to do that? Perhaps when you’re frustrated because you can’t find a job?

So its possible that businesses, regardless of offering, are getting poor ratings because customers are being asked to rate the app at a time when they’re unhappy.

You can test this hypothesis quite simply – flip the scenario and ask your customers at a time when they are happy and see what happens.

With job hunting the most positive times would either be when the jobseeker gets the job or when they’ve just applied for a job. Unfortunately neither were options for Jobsite due to process and technical limitations, so we theorised that the next best time would be after using the app on three occasions. With the high abandonment rate of apps, we reckoned if you’re using it for the third time, you must be finding it useful.

So we introduced a timed pop up – see example image – inviting them to feedback on the app.


The big reveal

The results?

The app rating jumped from 1.5 to 4 out of 5 immediately!

We hadn’t improved the app in any way, we’d just captured the views of the silent – and happy – segment, who until that point had had an obstacle strewn path to review the app.

One of the great things about having to resubmit the app to the App Store following a change is that the ratings displayed are for the new version. So any positive impact is felt immediately, as you’re not having your average score dragged down by all the previous low scores.

It’s worth noting though, you’ll still get the bad reviews – the product still needs work – but now they’re just balanced out with the good ones. In the meantime you’ve built strong social proof in the App Store and a better star rating equals more downloads.

So what are the lessons learned?

For a start, assumptions can be dangerous. When we’re pushed for time, it’s easy to look at ‘facts’ and take them as gospel. The star ratings said the app was no good, but your future business planning cannot be determined by hundreds or thousands of bits of customer feedback distilled down into a simple 5 star graphic. You need to go a step further and unearth richer data to make informed decisions.

You’ll come across roadblocks – such as an industry working a particularly unhelpful way – so you’ll need the resolution to not accept things as set in stone. You can either work out how to disrupt that industry (tip: focus on what would really make the customer happy) or you can perservere and find a way to work the system to your advantage.

This all takes time, time you perhaps don’t have. Consider this though, do you want to get lots of things done and maintain the status quo or do you want to get less things done but do them really well and shift the needle?