Anatomy of a scam

A (smart) friend of mine sent me this Whatsapp the other day.

whatsapp scam

It looked like a scam to me, but I was interested in understanding a couple of things:

  • how smart people get taken in (i.e. what marketing/persuasion tactics were used)
  • what benefit the person who set it up gets

As I went through it I had discussions with my wife who managed to decode the tech stuff behind and (spoiler alert!) find the culprit.

Marketing persuasion tactics

The initial message is good:

  • they use the official logo
  • the domain looks legitimate (but actually it’s a subdomain of
  • it’s short and punchy
  • it contains the magic word “FREE” (twice)
  • Urgency (as Cialdini says) is a great method to encourage people. (though I think they’ve tripped up here: “Hurry up! Collect your FREE voucher” is not the words of a respected retailer giving away R5000)

Unfortunately the site is now down, so you’ll have to look at these desktop screenshots to puzzle through the rest of the process.


Screenshot 2020-04-18 at 10.46.59

The page you click through to is on the left and the Woolworths homepage on the right. If you compare them you can see quite distinct branding differences (i.e. the left hand version doesn’t have great branding: small logo, lots of colours, “cheap” design). The image of the shop on the left hand version is also a bit strange, it looks like someone’s taken a picture and it doesn’t look like they’d put a graphic designer on a promo that would be likely to cost them a fortune.

Step 1:

There is a countdown from 85 remaining vouchers on the page (again Cialdini’s principle of scarcity). It almost seems believable that Woolworths would be giving away big vouchers (especially in the first days after lockdown). The precision of the 2 day expectation of when you’ll receive the vouchers is a nice touch, it sets expectation and makes the offer less nebulous. Having to jump through some hoops to qualify also adds trust. It show it’s not some lame freebie, one actually has to work for it.

Step 3:

Step 2 was an age verification, another trust marker to prove your eligibility. Step 3 gives you more hope if you’ve never won a voucher previously, that they’re using an algorithm to ensure it must be your turn this time.


More Cialdini magic here, some social proof to cement the fact that you’re definitely going to win as evidenced by these real people (or three people from the Smith family). You can’t click on any of the profile links to check that they’re real, but having them there as proof is enough for most people.

Unfortunately this is where the wheels fall off a bit. The tool they use is a template that they’ve used for other offers and they didn’t bother to change this offer for a free pair of shoes to the Woolworths voucher.
Technical stuff

If you view source (and are cleverer than me) you can find out there is a script that does most of the hard work:

There are a couple of sites referenced and if you do some digging through domain registrars you’ll find a clear link to this guy in Delhi:

Screenshot 2020-04-18 at 11.43.24
It seems that he isn’t a developer and is using a script that you can buy for perpetrating scams like this (Google voucherf0c9.js if you’re in the market for some low level scamming).

He has Google Tag manager set up and a number of services to track events, generate random social profile pictures and an AI tracker

It seems that it is set up specifically for mobile and has a load of affiliate links/cookies on it (+/- 134 cookies!!)

You can see an example of another promo/scam that they had on their site below.


The big question, why do this?

People want to get a great/FREE deal so they’re motivated to fill it in, but what about the person that is perpetrating it.

My assumption is that they dump a whole load of affiliate cookies on your device, you share it with all of your friends so there is a large viral spread of this. Eventually someone will purchase somewhere and there is an opportunity to make some money off you.

Simple and effective if it spreads far.


I had the pleasure of receiving a new version yesterday, as you can see exactly the same template:

The power of nudges

I recently upgraded my Balsamiq subscription and I thought this was quite a good example of what I mean by a nudge:

  • a consistent always on reminder of the payment option and the benefit that you’ll receive.

Screenshot 2020-03-05 at 11.42.52

I think it works best with industries like SAAS where you are on the site fairly frequently, but paying isn’t always obvious, and as the site owner you want to keep it top of mind,

I’ve written previously about when we started testing nudges on Now Novel, and since then we’ve broadened out the complement so that they drive over half of our sales.

We’ve tested a number of different areas where it is possible to reinforce benefits at different junctures :

  • modals that cover the screen
  • text links near your profile/sign-in
  • inline boxes and
  • top of the page persistent strips

I wanted to share some of the results that we had with a test that we were running. This was at the end of the freemium section where you have created your character. You have seen the value of the product at this stage, have some momentum and increased motivation,  therefore it’s a good point to encourage purchase.

This is the control/default option we had been showing:

Screenshot 2020-03-21 at 16.31.06

and we tested two variants, one with a focus on a testimonial:

Screenshot 2020-03-21 at 16.30.16

and the other with a focus on the benefits that people said they received from the service (so reflecting benefits back at them that we learned from surveys)

Screenshot 2020-03-21 at 16.31.45

We measured the impact of this by visits to the payment page (a useful proxy metric for progress to sales) and sales conversion. We had over 90% significance on both variants for sales conversion, but a higher uplift on the testimonial version. This is the one that we’re rolling out as our new winner.

Some next testing opportunities:

  • make the review more obvious (star ratings, who powers reviews to add more 3rd party accreditation, “What our customers say” to label it effectively)
  • combine some more of the benefit messaging into the testimonial version
  • can we make it shorter, easier to scan and punchier?
  • Is “Upgrade” a good call-to-action or could we use something like “Get access” (upgrade sounds like you’re spending a lot of money)

Headspace’s cancellation funnel- teardown

My Headspace subscription was coming to an end and I wanted to cancel . As I went through their cancellation funnel I wanted to look for dark patterns and see if there was anything I could learn from what they had done.

Overall I thought it was OK, but I think trying to rescue a cancellation at this stage is hard work. Even if you do, without finding a method to fix why people were cancelling in the first place (through some sort of re-engagment/re-on-boarding) would be quite difficult.

Step 1:

Screen Shot 2019-05-20 at 12.05.32

I thought this was pretty good, a good graphical representation of the features that they offer (having never used Headspace for kids I’m not sure why it got such a high placement, but maybe that is a key feature for most people). I am surprised that these are all feature driven, and not benefit focussed  (I’d expect them to be focussing on improved concentration, reduced stress, better sleep etc). The green “Never mind” is a strange call to action, I’m not sure exactly what it should do/why that is the exit option.


Step 2

Screen Shot 2019-05-20 at 12.05.45

This is an interesting screen. “Changing your mind” is a good turn of phrase, especially in the context of what meditation can offer you. The first two (customer support and payment plan changes) are quite mundane.

Helping you find time to meditate is a great option. We find that it’s a key challenge for people wanting to write a book using Now Novel too. Self directed long projects are challenging for people to commit to. They don’t have an idea how long they’ll have to do it for (which is scary) and they’re unsure if they have the gumption to stick with it. In addition watching Netflix or scrolling Instagram is a lot easier after a day working. I would expect that this strategy would be better employed when onboarding rather than at this juncture. People have low motivation when they’re cancelling, so asking them to try and start a time saving process may be difficult.

I was thinking if I was looking at this first screen through a different lens how I might test a change. I have no user feedback so it would be entirely based on my perspective (so it probably would be 100% wrong :)).


I have put together a very rough wireframe.

Here are my thoughts:

  • starting meditation can be hard, so reflecting that back to them and explaining that they’re not alone with the challenge can make them more receptive.
  • Focussing on the benefits, not the features
  • The marginal cost of one extra month for someone is so negligible in terms of cost, that you might as well try offering it
  • Getting people to start onboarding (or speak to customer service) may resurrect them
  • Putting in as much social proof as possible (60 million users, high ratings) may convince people that its them and not the service
  • you could be sneaky and put the cancellation in a text link, hide it elsewhere or make the button text something like “I’ll quit please”, but I don’t think that kind of dark pattern UX helps you in the long run.

Using data science to get insights from reviews

Review Analysis Wex Canon (1)

Machine learning and AI are frequently used buzzwords. In terms of application I’m not sure whether we’re in the killer app territory yet. I read someone on Twitter saying something to the effect of most AI can be replicated with some good SQL queries (which is a nice soundbite).

From a conversion perspective I use Eyequant quite a lot. It mimics heatmaps of user interaction from a design (either on live websites or on heatmaps/wireframes) so you can get an idea of how visitors will perceive a website and where they will focus on the design.

From their website:

“EyeQuant fuses leading neuroscience research with AI to accurately predict how people will react to digital designs.”

I have been working with a data scientist on a couple of research projects to try and parse reviews to get an idea of underlying perspectives. “Text mining” of reviews is not something unique, this article goes into some detail about how to use R to extract the data, but more importantly how to display it so that quantitative insights can be gleaned from the voice of consumer.

Here are two examples of how I’ve used it:

  1. A company was selling the same products that Amazon stocked and my thoughts were to understand both the reviews and the questions asked by customers (as Amazon has such a volume of feedback it made this process easier to look at Amazon rather than on their site). One of the key products looked at was a bundle. If I use the example of a keyboard, computer and mouse I was surprised to see that most reviews focussed on the mouse, rather than on the core product. This lead to testing exposing more details about the mouse as we knew it was such a focus for visitors.
  2. The other was also to look at reviews for a high-tech product, but to understand what features people were most interested in. Once I found out what the focus was this helped to prioritise what to surface more and how to modify the architecture of the page used to promote this.

These are some pretty straightforward applications, but its interesting to understand that there is a wealth of data available, and if we can learn to unlock it efficiently we can be more efficient.


How jobs to be done can aid your first user experience


I’ve drunk the jobs-to-be-done Koolade and there is no coming back. Jobs-to-be-done is a framework which allows for uncovering customer requirements in more depth, so that you can ensure what you build maps to what customers want. through the process you’ll conduct open-ended interview to understand:

  • what solutions customers are currently using
  • how they feel about them
  • what their ideal solution would do (even if it doesn’t currently exist)
  • what job they would employ your product to do

Your role is to then take all the responses, annotate and create high level requirements from this. We’re looking to understand what job the customer is looking to “hire” your product to do. Its not asking your visitors for product specs or to design the product for you, they don’t know ( the apocryphal quote from Henry Ford fits here: “If I had asked people what they wanted, they would have said faster horses.”)

However ensuring that you’re solving the right problem and building solutions to help with customer needs means that you don’t waste resources, saves time and results in happier customers. I’ve used this framework across a number of industries and it is remarkably robust.

I believe jobs to be done are more effective than personas. Most personas that I encounter seem cliched. A JTBD allows you to stack numerous jobs on a single person irrespective of their demographic or psychographic background. Others would say that a well sketched out persona should already include the JTBD and the additional detail in a personas allows you to have more empathy for the subject. I suppose it’s a matter of personal preference.

A colleague recommended some helpful books recently about Jobs to be Done:

  • Cindy Alvarez’s Lean Customer development
  • Alex Osterwalder’s Value proposition canvas (for the second time)

both of these are not just theoretical models about how to get insights, but offer actionable step wise processes to gain insights.

A new project we undertook was to use JTBD to aid onboarding and create a better first user experience. We have one onboarding sequence for everyone, irrespective of their requirements. As you can imagine this isn’t perfect. Once we outlined the key jobs that were required we went further with a survey to get even more insights. We asked a number of questions, but a couple were useful for this area:

  • finding additional challenges
  • finding satisfaction points

We could then segment this based on paying vs non-paying users as well as the core job they were interested in. Once this was analysed we could find insights that would match the specific job with the best part of the product as well as the additional challenges they had with parts of the product that help with these requirements. This would aid them and give them a more tailored first user experience.

More testing and development needs to happen in this area, I will report back on results.

Feature request: Intercom analytics

webaroo-com-au-shr_Xn8S8QU-unsplashWe’ve been using Intercom for Now Novel since 2012 and have watched it turn into the behemoth it currently is. It is a great tool that helps us communicate personally at scale.

Intercom doesn’t leave a lot of extra customer value. They know the value of their service and what people’s propensity to pay is and are not afraid to map those two quite close to each other (i.e. its not cheap :)). This could be an additional service they offer that could add more to their offering.

If I look at the tools that we use to get insights from our customers (mostly Hotjar and Google Analytics) we don’t have a tool that focusses on users. Google Analytics is a great tool, but it’s a little hamstrung by being session rather than user based. Sessions are interesting, but on a user level you get a lot more of the user journey and see where they interacted with what portions of the site on an individual level.

We’ve already set up Intercom already with a lot of information that would be great to use to understand our users further. Our Intercom integration has specific questions that are set up at key points in the customer journey to message to customers. This means that you’ve already got the tagging part of the equation sorted out. Intercom prompts are usually set at key parts of the customer journey too (e.g. end of onboarding, start or at payment junctures) Being able to interrogate these would be very helpful. We have already tried to use this before to look at the data, but it isn’t possible to look at it in an easily accessible manner. You have to go into each user and manually look at their progress, you can’t e.g. export a csv with all of a user’s activities.

I’d ideally like to be able to look at a tool that would allow me to:

  • look at users over a specified time period
  • who have achieved a specific aim or goal (which would be one of my triggers) and
  • examine the data around this criteria

The data around this criteria could be other goals, visits, length of time, communication acted upon etc. This starts to be even more interesting when you can import and export this information into other tools (e.g importing my Google Optimize test variation into Intercom for segmentation on how a specific test performs or a specific segment I have created in Intercom that I can start to view more information about in GA).

Intercom must have thought this, so when they need an early beta-tester, count me in.

Books read 2020

  1. Lean customer development- Cindy Alvarez

My first completed book of the year was very useful. I’ve read a fair amount in this area (the book references the Lean Startup and Steve Blank), but this was a useful addition to the canon. The best thing about this book was the  practical /tactical method it was written in. It taught you the theory, but then it shows you exact methods on how to implement your customer learning process and how to quantify and use your learnings. There is immediate practical value I am taking from it to use in my process.


2. The big nowhere- James Ellroy

I have read (and watched) LA Confidential and am a big Raymon Carver fan, so I was prepared for the hard-boiled cop fiction. It is quite unrelenting however; brutal murders, wolverines, communists, crime bosses. It doesn’t let up for 400 odd pages and goes through numerous twists and turns. The heroes aren’t really apparent and the main characters all have their own struggles. However it is definitely worth a read, the perspective on marijuana and homosexuality in the 40’s s especially interesting knowing how in modern day California they aren’t even an issue.


3. Grace Jones- I’ll never write my memoirs


Thought this isn’t the cover, its hard to talk about Grace Jones without talking about her image (and my favourite of her images are the ones she created with Jean-Paul Goude, he talks about this one here).

I am a big Grace Jones fan; I listen to her music, I watched her documentary so this was interesting to hear straight from the horse’s mouth. Some of it was a little repetitive; pushing against the vagaries of the established music system for 40 years must be challenging, but can also be much of the same. There were flashes of her personality that came out that sound amazing and tiring (if you had to behold them). A perspective of a unique artist.

Builder’s Warehouse- perplexing purchase flow

I needed to buy something from Builder’s Warehouse on the weekend and I found their purchasing flow so perplexing that I had to document it.


Step 1 of the process was that I wanted to buy an online only product:Screenshot 2019-10-26 at 10.12.16

A couple of comments:

  • Why would you have the option to “select a store” if it is an online only product? It only confuses people when they wonder why there is store availability for an online only product and they question whether they should be shopping from you.
  • “Sign in to buy” is a tough call-to-action. Someone has to have a lot of motivation to get that product to sign up to have the pleasure to give you money. Most visitors would disappear at this stage if they didn’t really love/need that product.


Step 2- signup

I am a masochist (and I was intrigued at what would come next in the process) so I clicked through.

Screenshot 2019-10-26 at 10.12.02

More comments:

  • Adding in additional non necessary fields is a sure-fire way to decrease completion of your form (especially if you have to sign in to buy). Why do you need my birthdate? Are you going to steal my identity with my passport/ID number? These are not things I want to share with a brand…
  • The chutzpah of the news form. No option to not receive any news, but an opportunity to choose both email and phone spam. I know this form certainly doesn’t adhere to GDPR, and although I’m not an expert I think it falls foul of the Protection of Private Information Act (and as a consumer its a big conversion killer)

Step 3- registration complete

Screenshot 2019-10-26 at 10.14.16

Now I’ve gone through all the hoops to try and purchase my product, but once I’m registered it spits me out on a registration confirmation page and invites me to go back to the homepage (where I have to find my product again). It would have been easier to save the consumer a couple of clicks and put them back on the product they were looking for.

All in all a very confusing process. If anyone at Builders Warehouse reads this, give me a call, I can help 🙂



Rules for split testing for startups


In a startup you have limited resources and you want to get as much bang for these resources. As a result I wanted to recommend some rules that have worked for me:

  • Err on the side of bold testing

Its nice to know exactly what was the underlying cause of the uplift from your test, but if I have the opportunity to run three tests I want them to give me the biggest uplift, not the most knowledge. This means that you’re opportunity for failure will be higher, but when you win it’ll translate to a lot of wins. Meek testing (e.g. button colours, CTA buttons) might work if you’re Google or, but with limited traffic and resources you want to focus on things that are going to move the needle for you.

  • Make falsifiable hypotheses which will help you learn even if the test doesn’t win

As a corollary to the above you want to learn as much as possible from every test. I like the hypothesis format:

As a results of evidence

We believe that the following changes will result in this effect

We will measure this through this metric

If your hypothesis gives you a learning then it will make it easier for you to create better future tests and iterate on your testing.

  • Each test should run minimum 2 weeks.

If you run your test for less than 2 weeks you won’t get enough traffic and you won’t understand the experience that people have of day parting (i.e. your test results may skew to how people behave on  a weekend or week day). This also depend on the amount of traffic you have, you may need to run a test for months on low traffic pages (which then begs the question, why are you bothering to test this page, see “test important things” point below)

  • If after a month the variant has a significance level of less than 65%, kill the test

You need to keep a decent testing velocity, if the variant is only slightly better than the control then it doesn’t seem like its going to be a winner, and its not worth continuing to test and waste the slot.

  • Significance level of at least 75%

All the tools tell you to test to 95% confidence. If this is your payment page or a crucial page then I’d agree that its worth testing to a high significance, however if it is smaller pages you have the challenge of getting enough traffic and having a number of things to test in your pipeline. In that case using a lower significance level is justified, especially if its backed up by:

-performance over time

Its also interesting to look at how the test has performed over time. If the variation has always been winning then you can make an assumption that it will continue winning.

– secondary metrics/micro conversions

  • Aim for micro-conversions (i.e. next step in process) as opposed to sales/far away goals

Conversions and revenue are your obvious goal, but often you don’t have enough traffic or enough conversions to get statistical significance on these. Micro-conversions can help here; clicks on “Add to basket”, reaching the basket or the next step after the elements you’re testing can add further colour to the picture and will be easier to achieve significance on than metrics further down the funnel.  These can be proxies for conversion especially if you have a few of them to review in addition to your main metrics.

  • Test the most important things

Don’t test your about us page and pages that have a limited impact in the purchase journey. Elements like sitewide navigation, signup, checkout etc impact everyone on site and so will have a much larger opportunity to move the needle when it comes to achieving meaningful increases.

  • Have a roadmap

Having a number of tests planned for each slot helps, then you don’t have to go back to try and puzzle a new test every time you conclude a test. If you invest the time in having one or two tests schedule for each slot then your process works more smoothly.

These are some elements that have helped me keep a healthy test trajectory with limited resources.

Quick customer insights hack


Retention is a big initiative at Now Novel at the moment and I’ve been looking at speaking to paying users to understand motivation and how to improve our onboarding process.

I have been sending the below message through Intercom to paying customers of our non-coached programs to try and speak to them. They get the opportunity to schedule a meeting with me on Calendly and then we chat.

Screen Shot 2019-06-07 at 08.44.15

  • Its been very effective for a set it and forget it process (and more effective as an email than as an in-app message).My agenda is very loose for these calls:
    • I introduce myself
    • understand their challenges/progress
    • explain the product (while trying not to upsell too much 🙂 )
    • let them know they can always get in touch


    Its been very useful and I’m going to continue doing it. It really connects with people that you’re making the effort to speak in person (which isn’t particularly scaleable, but its nice to make the offer anyway). You understand where the gaps are in your onboarding process (we have a couple of features which aren’t as obvious to find, and it seems people aren’t finding them), and you can get feedback from people who are motivated but are finding inconsistencies in the process.

    It’s also fun speaking to a wide range of different people with different backgrounds and motivations. We had a lawyer who purchased in Cape Town/up the road from me, so I went in and spent some time with him in his office rather than speak on the phone.

    In terms of a continuous feedback loop, this is an easy and recommended process.