Grace’s 2024 Posts

People always say “learn to speak finance’s language”, but what does that actually really mean?

In my new article – link in comments – I set out a LOT of super practical things you can do, suggested to me by real life CFOs and finance directors.

⭐ Performance and retail media ads help customers to find you online, mainly doing an availability job. So, present them as operating expenditure – something that’s used up straight away – and show you’re minimising costs.

e.g. “Amazon is like a machine, we’ve got to spend about 20% of the retail price on ads there. You spend more, you don’t get much upside, you spend less, you see sales drop off. So, it’s like an OPEX”

⭐ Brand is different because here you’re buying something that pays back later. So present it as a CAPEX and show you’re maximising the payback and minimising risk.

There’s more in the article, including how to use scenarios to give finance a sense of security, the metric that allows your brand campaign to be compared to R&D or some other investment, and the move that turns your plan from risky into a safe bet.

Read the article here: https://magicnumbers.co.uk/articles/villains-heroes-finance/

Econometrics is the most reliable method for understanding the results of past marketing activity.

It gives you a cost-benefit analysis of everything you’ve done, so you can see what’s best to do next.

This interview with IPA (Institute of Practitioners in Advertising) and LBBonline – Little Black Book is a simple introductory guide, including information on:

• Why Martin Sorrell called econometrics the “holy grail”
• What the benefits of doing it are
• A really simple example of how it works
• Why it’s better than attribution, especially last click
• So if it’s something you’ve always wondered about have a read, follow the link here:

https://www.lbbonline.com/news/answering-burning-econometrics-questions-with-dr-grace-kite/

The great X-scape?

As Elon Musk’s X has steadily reduced features and moderation in the name of free speech – and Musk has become more involved with Donald Trump’s presidential campaign – it seems that more and more users are becoming ex-Xers.

The chart shows the scale of the migration away from X.

Daily active users of X (the number of people actually opening the app and scrolling through it) has been on a long decline. It was down 30% in the UK by September this year compared to May 2023.

As events have transpired this week and Elon Musk has been appointed co-head of the Department of Government Efficiency (Doge), this decline has accelerated massively.

It is a loss to the media stack, particularly for B2B businesses that valued being involved in the professional conversations that took place on twitter.

So where should those businesses go now? Well, it depends on where the users are going!

Mark Zuckerberg’s Threads seemed like an obvious choice after its initial extreme growth back when it was launched in mid-2023. There were over 100m new members in its first week! But, Threads’ growth in number of users has stalled hard since.

By contrast, BlueSky, launched by twitter founder Jack Dorsey had a very slow launch, with essentially your-name-on-the-guest-list required to actually get on it.

Right now, though, it looks like BlueSky is winning out in having more appeal to X-iters looking to find a social media experience similar to what Twitter was before Musk’s takeover.

So, if it reaches a critical mass, it’ll be more able to step into the rather specific breach left by the little blue bird.

Have you had thoughts about ghosting this X?

Is influencer marketing right for you?

The chart is from a big new survey by YouGov that asked over 10,000 respondents all around the world all about influencers.

It shows the topics where influencers are particularly, ahem… influential.
Around 60% of people find influencer content on travel, fitness, tech, and fashion useful. So, these products are good contenders for a strong effect of sponsored content on sales.

At the other end of the scale, the topics we don’t want at dinner parties are also the ones we avoid in conversation online: Politics and religion content is seen as the least useful.

But be careful, because culture matters. For example, in Indonesia, 86% of respondents find travel content useful compared to only 36% of Danish respondents.

It’s all part of emerging learning about how best to use influencers.

Other recent gems include a report by The Drum that Gen Z audiences trust influencers more than celebrities. And one to treat with some caution from the influencer benchmark report, which shows the average payback from influencers is $5.78 per $1 dollar spent.

If you’re not yet ready for big business brand-building, but worry you’re too reliant on performance marketing, influencers, just like online video can be a low-cost route to a new phase of growth.

Find out how to tackle it in my super practical, accessible, and no-nonsense Scaling Up Works course. Sign up by going to the courses page at the top of the screen or chat to Imogen for group bookings and questions via [email protected].

It’s a slippery slope…

Moving money from traditional budgets into search ads on retailers’ sites, is like paying the rent out of your savings. It can only lead to poverty ahead.

It’s because retailers, and especially Amazon, control such a large share of product buying traffic, you have to be there if you want your product to be available to buy. This means they can charge what they like.

The chart shows that, aside from a COVID blip, it’s costing UK businesses more and more of their revenues and profits, each year.

The slippery slope comes because the only way to avoid continuing to pay as prices for this type of ad go up is to invest into your brand.

Because if people want your product before they get to the search page, any set of search results for the category which doesn’t include you is “wrong”.

And auction algorithms, which value relevance as well as advertising revenue, will allow you to be there for a reasonable price.💰

So don’t plunder your TV or out of home budget to pay retail-media rent in 2025. Instead, make the case now that this must be paid for out of the distribution or merchandising budget.

My latest article, link in comments, is all about this challenge, and how to navigate it in the coming years.

🚩 If you want to avoid paying for online ads that you don’t need and get the balance of brand and performance right, you should join our course Scaling Up Works.

If you want a bigger marketing budget you’ve got to be able to talk turkey🦃

Marketers who measure and report return-on-investment are 1.6 times more likely to get a bigger budget.

Put yourself in your CFO’s shoes. He or she isn’t a baddie, just a practical numbers person that wants to do the things that’ll make the company grow and avoid taking too many risks.

The job is to adjudicate on an inbox full of 100 different things the company might buy, from machines in the factory, to an R&D hire, to a better fleet of delivery vans.

And every time something is approved it comes out profits and makes that all important P&L look worse, at least for a time.

The decision to approve is about two things: The likely payback and the chances that the person asking the money will manage this expenditure well.

Measuring ROI ticks both of those boxes because it not only quantifies the benefits of marketing, it also shows that you’re doing the work to learn and improve how you do it.

🚩If you want to learn more about advertising ROI, including how to do a rough estimate for yourself, how to commission analytics for an accurate read, and how to avoid pitfalls, you should join our Data Works course.

There’s a subtle effect of being on TV – via the Google auctions – that most people don’t know about.

It’s no secret that pound for pound TV often comes up in econometrics as one of the best channels for delivering visits to your site and converting them at scale.

So, its no surprise that this finding was confirmed when ITV carried out 30 rigorously designed regional tests. The chart shows that, on average in these experiments, running a TV campaign brought a 23% increase in site traffic and a 33% increase in conversions.

The tests also revealed a less obvious effect, working in the background.

They showed that TV actually influences the paid search auction.

In fact, a TV campaign makes you much more likely to appear at the top of brand search term auctions, to the tune of an extra 41%.📺

This means businesses that buy TV are more likely to appear in the top slot on Google when people search for them, and they pay less, per click, for their customers’ journeys through the search environment.

The success of media channel is often considered in a silo. Did it drive an uplift in conversions or awareness or some other similar KPI? Indirect effects and synergies are rarely fully taken into account.

A good measurement framework will set you up to understand both direct and indirect effects of your marketing mix.

Two thirds of marketers say their business doesn’t fully understand marketing.

That’s scary if it’s true, but data in charts doesn’t always mean what you think it means.

Maybe the truth behind this particular pie is that the business does understand, they just don’t agree.

Marketers – and particularly those that understand the benefits of brand building – try to rebrand advertising spend as “investment”. When you buy it, they say, you purchase an asset – positive associations in people’s minds. That asset pays back over time in future sales.

The trouble is, advertising quite literally *is* a cost on the P&L.

It’s not an asset on the balance sheet, like the laptops, desks, and factory machinery. In the accounts it doesn’t register as something that’s owned. It can’t be sold.

Telling someone senior who understands the mission to get profit in the actual accounts that will actually be submitted to companies house that it’s investment might actually not be helping.

Maybe it’ll confuse that person, or it’ll make them think ‘stupid marketing people just don’t get this business’.

So maybe, just maybe, this is one where it’s not worth fighting a battle over the terminology, but instead refocussing on explaining the benefits of incurring marketing costs.

And if you want to do that using killer charts that really do represent the truth of the matter, check out our training course Data Works.

Want to find out what Les Binet is getting goosebumps about?

Link to our Pubcast video here: https://lnkd.in/eT8mYk9F.

That’s right, it’s a podcast, but from a pub.

It’s got beer! It’s got giggles! And yes, it’s got lots of marketing effectiveness genius.🍻

Great acts of brand building now have to be made up of a lot of little things.

Because there’s no going back. The platforms that people scroll through, only catching a glimpse of your ad, are here to stay.

And the generation that’s grown up with content that suits them exactly isn’t going to start watching please-all programming in their millions.

It means that smaller reach per placement and smaller attention per ad is not going away.

But all hope is not lost.💡Because the modern media stack includes numerous ways of communicating that have been shown to work well together.

If you can get “matching luggage” – ads that look and feel the same, but are also tailored to the platforms concerned, you can get the benefits of a “lot of littles” campaign.

Because of synergies, going from using 1 channel to using 5 brings 35-65% more ROI and double the effect on brand metrics.

No single media channel gets everyone for 30 glorious seconds, but together, the stack delivers a critical mass of smaller exposures, without getting boring… and, if you can manage creative and media so it all feels consistent, it works.

My latest article, read here: https://lnkd.in/eQJrapuM, is all about this route ahead, and what you need to know to use it in 2025.

Free and simple online tool to help you use Econometrics/MMM!

I love Econometrics/MMM. Not in that cringey way you see on CVs, where people claim a passion for work to impress a future boss. But it in an unabashed, geeky, I know it’s not cool but I can’t help it, way.

It’s just a brilliant tool for learning how the world works using data. And it solves all the well-known problems with last-click attribution. And it doesn’t need 3rd party cookies.

I want everyone to be able to use it in 2025, so, with the team at magic numbers I’ve made a tool that will help you use it during this Autumn’s annual planning.

It does 2️⃣ things:

1. Is for smaller businesses – the tool tells you if you’re ready for econometrics/MMM, by assessing whether you’re spending enough to get a good read with this method.
2. Is if you’re already using econometrics/MMM and experimenting with new media channels. The tool tells you how much you need to spend on that channel for econometrics/MMM to give you a good read.

Give it a go here: https://lnkd.in/eycDxPNM, and pop your thoughts and questions in the comments. We didn’t write a manual yet, but we’re here to support you as you get stuck in.

People trust influencers.

The set of pies below shows that around three quarters of people who follow creators that promote products or brands like the recommendations they get.

They believe the influencers they follow are on their side.

That they promote affordable, reliable, and good quality products, and that they do it because they’re looking out for the best interests of their followers.

And, just in case you were wondering whether c.75% trust is good for product recommendations, I can confirm, it is.

• It’s way better than celebs see. The proportion who trust celebrity endorsements is a paltry 1-3%.
• In fact, it’s pretty close to the degree of trust people place in a recommendation from a friends or family member at 92%.

It means that these people offer your business a modern way to access the benefits of positive word of mouth.

And that means you have an option for brand building in 2025 even if your budget won’t stretch to luxuries, or your audience is young and not watching TV any more.

If you want to find out more about brand building online, including strategies that won’t break the bank and what’s worked for other businesses, check out our training course Scaling Up Works.

Is your budget set at the right level this year?

We’ve done the work to establish a good rule of thumb for how to set your advertising budget relative to the size of your business if the objective is to maximise ROI.

The chart is from the ARC database – a joint initiative between the IPA and magic numbers to bring hundreds and hundreds of econometrics findings from 6 econometrics shops together.

It shows the optimal % of turnover you should devote to advertising if you’re an online business and you want to maximise ROI. It’s about 5%, and for online businesses the ROI is higher at that optimum than it is for bricks and mortar.

But do you know what else we found? It’s that the same finding was there in 2 other reputable sources – in Nielsen’s ROI study, and in an investigation by Paul Dyson. All three studies landed on spending 5-10% of turnover on advertising for the highest return on investment.

This chart is taken from our course, which is all about putting your money where the magic is. The right size budget and the right mix of brand and performance marketing.

If you like using a benchmark, meta study, or rule of thumb to make the right decisions *and* successfully sell them in to your bosses, learn more on our Scaling Up Works course.

Measurement’s a mess!

Nearly two thirds of advertisers in the UK and USA still rely on last click attribution to optimise campaigns.

And that’s not just a little bit sub-optimal. It’s very costly.💰

If you choose using last click, you miss channels that work earlier in the purchase journey, even when they are the best thing for you. The middle pie is data from Analytic Partners – it shows that with last click, you miss out on 35% of possible sales achievable for your budget.

We get it, it’s understandable, last click is sticky.

Perhaps you’ve made decisions using it in the past and it would be a climb down to change them now.

Or, maybe it’s just that the numbers are always there and changing for something else would mean less or slower access to intel.

But it gets the answer wrong.

It’s no wonder the right-hand pie shows 40% of marketers are worried about measurement.

Killer charts are powerful things. Here’s an example of how to make one.

It comes from VCCP and Cadbury’s IPA effectiveness awards Grand Prix winner in 2022, which we at magic numbers helped to author.

It starts off as a humble line chart, showing the change in Cadbury’s penetration over time, but fresh out of Excel, the chart doesn’t really show anything. It looks flat and uninteresting.

But give it a 4️⃣-step makeover, and this pedestrian chart turns into a killer.

1️ Make the change in penetration abundantly clear to the reader by truncating the y-axis – that means chopping off the bottom of it.

It’s fine to do this for line graphs, because the important thing is the change over time, not the height of the points. It’s an absolute no-no to truncate the axis on bar charts.

2️  Annotate the chart to make it even clearer what the reader is supposed to see. The turnaround in penetration coincided with the campaign launch. Great! Write it on the chart!

3️  Take stuff that’s not needed out. In the frame above, there are no grid lines, no axes, and no label on the x-axis. Who needs a label to know that those things are years?

4️  Add stuff in that tells your story. The title is the message, the campaign period is highlighted in colour and impossible to miss.

Want conversions? Choose reach!

You’ve read that attribution isn’t that reliable.

But what to do instead? You know you want the type of online media that’s going to drive growth. But where to find it?

The chart shows that social is more likely to drive growth when:

• It’s got some new-ness. When it finds people who didn’t already know you, and features products you aren’t already famous for.
• It’s a video. Static ads are easier to scroll past and easier to forget. They remind people who are already thinking of you, but less often change people’s minds
• It targets video views or reach. If you choose objective clicks, you more often get people clicking who are using your ad as a navigational tool

Bar charts are the bazooka of data communication.

It’s because they’re the easiest way to understand data. The first chart kids get taught at school. But it’s easy to muffle the effect by making them look complicated, even when they’re simple.

Take this one from my 2021 presentation for the IPA. We’re looking at average return on investment from advertising by size of business.

This is a fabulous piece of data. Business size is the most important determinant of advertising ROI, so this chart is one you can refer to forever, to see if your ROI is better or worse than it should be.

But the starting chart is a mess. There’s way too much going on. A lot of “ink on the page” doesn’t add anything to the reader’s comprehension of the central point.

✂️First step is a FIERCE declutter. Remove the background, legend, axis labels and funky bar formatting.

And then add stuff. Put data labels on the bars and write what you want the reader to see in the title and annotation.

Tada! It’s now a killer chart worth a book mark for any marketing professional.

Is effective marketing like playing chess? I think so.

It’s something I explain in a conversation with Edward Cotton Chief Strategy Officer – CMOs and marketing directors should lay down a strategy, and then use measurement to check that it’s working and stay on course.

The conversation also covered:

• How marketers sometimes stray too far into simply reacting to every bit of data that comes their way
• The ever-expanding role of retail media
• The challenges that performance focussed businesses encounter when adopting brand building
• The importance of story telling with data

Highly recommended if you want to think big picture while staring out the train window, or walking the dog.

Communicating data well creates change. It really does.

The Long and the Short of It is famous, much quoted, and talked about. But what not many people know is that it actually changed behaviour in a big way.

The chart shows an indicator of short-termism in advertising planning. Up until 2013 when TLATSOI was published short-termism growing fast.

But afterwards, slowly, one advertiser at a time, we changed the planning horizon. The indicator of short-termism had a turning point and began to fall. It’s down a whole 10 percentage points since its 2014 peak!

It’s because of the stonking work done by the authors and publishers Les Binet, Peter Field, and IPA.

📈But it’s also because TLATSOI had a killer chart – the yellow and orange one with steps and spikes – and a killer stat – 60% brand, 40% activation.

A good bit of data, communicated well is powerful. It can absolutely bring change. Good change. Progress.

And you don’t have to be a super hero like Les Binet and Peter Field either. If you want to be able to make killer charts and stats that make change happen in your organisation you can learn how on our Data Works course.

Movie watching is forever changed since COVID.

UK cinema admissions are down by around a third compared to 2019, despite recent big-budget movie releases like Dune, Fall guy, and Mad Max.

Even the final Daniel Craig James Bond film and Barbenheimer underperformed vs. normal pre-covid hits like Star Wars: The Rise of Skywalker.

This isn’t necessarily cause for concern for individual advertisers’ payback from the channel – because cinema ads are generally bought per thousand admissions.

But it is a big loss to advertisers as a whole.

Cinema is great for telling stories, scoring highest in its ability to trigger a positive emotional response (Ebiquity). More than half of regular film-goers are positive about the ads they see vs only one third for TV (old salt).

Part of the problem is streaming.📺

In a cost-of-living crisis, it’s hard to justify spending nearly £40 to go out to watch a film, when, if you wait, you’ll be able to watch it at home on your own sofa, on your own TV, for a fraction of the cost.

And, if you’re a film-maker that has a streaming war to win, it’s very tempting to break the traditional rule of keeping new movies for cinemas only, and use them instead to drive subscriptions.

A lot now rests on experiments like Disney/Pixar’s recent Inside Out 2, which has been reserved for theatres for 100 days. These trials have to pay off if cinema is to win back its exclusivity over new movies, and regain its lost audiences.

For a big impact choose a small influencer.

The chart shows that return on investment from influencer marketing is highest where influencers have fewer followers*.

It seems strange, but it actually does make sense.

• Smaller influencers have fewer, more engaged followers.
• They talk to a higher proportion of them on a more personal level and even mirror their followers’ language
• So, each promotional post has a greater effect per follower and per £1 spent

This is a good step towards learning how to use influencers well, something that will be important for anyone wanting to build brands in the platform world.

• Influencers are well placed to produce the right creative: Tom Roach calls them ‘platform native creators’ uniquely able to match content to the platforms they use.
• And they’re trusted. In a survey by WARC’s nearly three quarters (74%) said they trust influencers to promote products in the best interest of their followers.

At magic numbers we often evaluate influencers in our econometrics and we’ve found them to be effective at driving sales over the long term.

All about geo-testing! If it’s something you’ve heard about but don’t know when to use it or how it works read on…

There are lots of situations where geo-tests work to answer questions about media, but the most powerful are when:

• You haven’t run a media channel before and you want a low-cost way to see if it works
• You think a media channel you’ve always spent on doesn’t drive many sales but it feels too risky to switch off completely
• You want to verify a test result from a media especially if it showed strong results e.g. Google or Meta

They work by breaking the country up into different regions, with one set of regions running the ads the other not.

Let’s say an online business want try out YouTube for the first time. They could split the country up like the map on the left of the image. They run YouTube in the blue areas, and they don’t run it in the pink.

If the campaign lasts for 10 weeks, then we might then end up with something that looks like the chart on the right.

Sales for the two regions match very closely prior to the test, then they increase for the region that has had the YouTube activity, before converging at some point after the test.

The difference between these two lines is the impact that YouTube has had.

Hooray! We have proven that YouTube works and quantified how well it works too.

One watch out: This was a simple explainer, but analysts do have to make a few decisions and judgement calls… so make sure you’re getting your test done by someone with no ulterior motive.

Moneyball media channels: The latest effectiveness stats reveal they are radio & podcasts, linear TV, and, would you believe it? Print.

⚾The movie Moneyball revolved around a strategy for winning at baseball invented by a guy called Pete Palmer.

He saw that, in choosing the same sort of players as each other, typical baseball managers played it safe, and in doing so, ignored players that would otherwise have helped them win. This opened up an opportunity for a brave manager to do something different and end up top of the league.

It’s a story for marketers as well as sports fans, because a marketing team that can pick media channels that work well but aren’t in demand has a distinct advantage.

The chart above reveals these moneyball media channels using the latest effectiveness stats from Profit Ability 2 by Thinkbox with Ebiquity plc and GroupM

• It has share of profit amongst the 141 businesses in the study on the x-axis, and share of total spend on the y-axis.
• Media channels that appear above the 45 degree line are fashionable. We’re spending more on them than we should given how profitable they are.
• Below the 45 degree line are moneyball channels, the choices that work, but don’t attract their fair share of spend.

In the movie, some of the best picks made by Brad Pitt’s character, Billy Beane, were older players who’d been prematurely benched. And the analysis in the diagram shows the same thing is true for media channels.

It’s radio and podcasts, linear TV, and, for the audiences that still read it, print, that are the moneyball channels of the post-covid era.

Read the full article: https://magicnumbers.co.uk/articles/what-the-latest-effectiveness-stats-reveal-about-moneyball-media-choices/#

Analytics findings should feel like Lego: Easy to handle, empowering, and usable for building stuff – stuff like revenue, profits and pay-rises.

But too often they don’t…. You get a deck full of 5 syllable words you’ve never heard of and slides that exist to prove the analysts are clever.

You get caveats and warnings about how not to use the findings, and never a straight up, clear, next step.

It doesn’t have to be that way…!

I’ve worked with WARC on a guide for non-technical people that want to commission MMM well and get the most from it. It’s got no jargon in it, but loads and loads of practical tips and help for normal marketing people who just want to do better marketing.

So if you don’t fancy having hierarchical Bayesian multicollinear endogeneity slapped on your desk at the end of your project, check the link: https://magicworks.training/wp-content/uploads/2024/04/Marketing-Mix-Modelling-A-How-To-Guide-for-Marketers-FINAL.pdf

📨 Email is the steam engine of marketing, it just keeps chugging on

For marketers, email has got a lot going for it: Everyone needs to check their emails, it’s cheap, and it’s not controlled by global tech giants.

But for customers, the prevailing wisdom is that it’s a bit of a pest. Clogging up your inbox with marketing messages when you’re just trying to answer the boss or set up a playdate for the weekend.

But no. Surprising as it seems, the data shows click and open rates for marketing emails haven’t wavered* since 2005. For nearly 20 years, a solid 30% of recipients open emails, and 3% click through to the website.

Now, obviously, email has its limitations. It’s never going to be a channel to hang a huge brand-building marketing campaign on, and you can only email people who gave you their email address.

But it’s also an example of something that, for marketers, is old but still gold. The best thing to do isn’t always the newest and the shiniest, and it’s only with data that you can find the right mix.

🏃‍♀️Measuring something changes the outcome.

That’s what this chart of the distribution of marathon times tells us.

4 hours is the milestone time to beat for many runners, and you can see it in the data.

Because so many runners really bust a gut to make it over the finish line before the 4-hour mark, we see that 3’58” & 3’59” are the most common times, after which there’s a huge drop-off.

Early in the 5th hour there’s no real incentive to push.

Measurement changes behaviour in marketing too:

• Pre-testing doesn’t always predict ad performance. Because people know they are being watched in the test, but not in the real world.
• Same story with survey data. People don’t say what they do.
• And just because someone “likes” something, it doesn’t mean they’ll buy it.

Want to hack brand building? Online video is your low-cost route onto the TV set.📺

The story usually goes that the rise of streaming has been a nightmare for advertising, because our TV sets have been invaded by high quality content that doesn’t get interrupted by ads. Brilliant for consumers, but a disaster for brand building.

But now something else has happened.

YouTube has arrived on the telly, stealing share from Broadcaster TV. YouTube now represents 4.8% of all viewing on the TV set, and 9.4% for 16-34 year olds.

These are ad-funded video services coming to the living room on the big screen with the default sound on. They bring a compelling short format that people like, and unlimited searchable content. Brilliant for consumers, and brilliant for brand building too.

And online video works for brand building. It has high return on investment, and long-lived effects. It must be true, 3 separate sets of research using meta analysis of econometrics found the same thing:

• Analytic Partners found online video has better short-term and long-term ROI than TV, and all video’s lasting impact is twice as long as non-video
• Mediacom/Wavemaker/Gain Theory found that online video has the same brand building effect as Print, OOH and BVOD, and comes with immediate data on how it’s working
• Meta found that online video drives almost as many sales as TV, with the best bang for buck of any channel

So, if you sell online, and you want to get the balance of brand and performance right, you don’t have to convince your bosses that an expensive TV creative and broadcast telly is the way forward.

Hopefully, you’ve learned something interesting here about balancing brand and performance marketing, but what drew you in? If it was the killer chart, or stat, or you want to learn more from the “how to” guide for marketing data you should check out my course Data Works.

Did you know YouTube has the highest reach of any ad-carrying platform in the UK?

Equal to Facebook & Messenger combined. And more than Google search.

🤯Surprising right?

Surely ‘googling’ is something that absolutely everyone does. It’s been in the Oxford English dictionary since 2006 after all!

But no, nowadays everyone’s on YouTube.

Your mum. Your grandad. Your kids especially. That’s where YouTube has the strongest hold, reaching 96% of children aged 8-12, and suggesting this trend will continue long into the future.

It’s an interesting one for advertisers, because we know that online video is a unique channel capable of brand building, with stronger long and short-term ROIs than just about any other channel. If you don’t believe it stay tuned, I’ll be posting the evidence shortly.

You can get started with smallish budgets, and apply test and learn strategies, with the immediate data you get back from the platforms – particularly useful if you sell online.

The marketing industry lacks data and analytics skills above all others, according to Marketing Week’s recent survey of over 3,000 brand-side marketers.

This is a 2.5 percentage point increase on the number from 2023’s survey, in which data analytics was already firmly in first place.

Give the 4Ps a chance? Clever headline, but I bet you’re sitting there thinking “my job’s just advertising and I’m busy enough, thanks”

I get it, but the truth is, the other 3 Ps – price, product and place – still matter🤝

Because if they aren’t all sorted, they undermine the effectiveness of advertising. And when they’re working really well, they can multiply the effects 10-fold.

It’s because advertising is a weak force, not a strong force. It isn’t capable of persuading people to buy on its own.

People just aren’t that easy to convince. To give a silly example, no advert can convince someone to buy dog food if their pet is a cat, and the principle is true even in less silly situations. There’s a huge set of academic studies that confirm it, many of them by Andrew Ehrenberg.

So even though you, in the marketing department, might not have power over the other 3 Ps, you need to understand them, and design your advertising plan accordingly.

Don’t air in the region where you’re not stocked in all the shops, and hold off spending until the product team fixes the font size on the labelling.

And bet your big budget when the price relative to competitors is just right.

Easter Eggs. They don’t make ‘em like they used to.

🥚It isn’t just that your hands have got bigger from when you were a kid, many eggs are smaller than they were, and they haven’t got any cheaper to make up for it either.

But there is one brand amongst the pack that’s bucking the shrinkflation trend. Where Mars, Kit Kat, and Minstrel eggs are smaller, Crème Eggs have actually got bigger!

They’re 15% bigger than they were at launch in the 1970s. And, even though they launched at 7p and now cost 75p, that price rise over the course of 50 years isn’t too much. It’s broadly in line with inflation.

Thanks Cadbury! Very generous of you.

Generous to all the kids and all their sticky fingers come Sunday. But also completely in line with the brand’s values and a demonstration of them living their positioning in how they act, not just how they talk.

Cadbury’s IPA Grand Prix winning advertising was all about generosity. The paper said it wasn’t just a pretend purpose, but something the brand really lives and breathes. This proves it true again!

Marketers take note. You’ve got to understand how all elements of the product, place & price are working for your brand, before you can get promotion (i.e. advertising) firing on all cylinders and working well.

Putting prices up is how 63% of CFOs prefer to manage uncertain times.

That’s what McKinsey found out in their Global CFO Survey last year.

And if every brand could put their prices up without taking a huge hit to volume sales, they surely all would. The CFOs of the world would be very happy.

It’s of course not quite as simple as that.🤯

We’re living in a world where inflation is slowly coming down, but remains relatively high – as high as 5% in food and drink according to the latest data released week.

So things definitely aren’t getting any cheaper. And shoppers are taking evasive action.

• People have been switching to own label. Sales of cheapest own-label products were up 47% last year
• Twice as many people were trying new retailers in 2023 v those in 2021. Aldi and Lidl are the biggest beneficiaries of this, with their combined market share now up to 16%
• And more people have been comparing online, according to Google searches

In this competitive environment, brands will have to work extra hard to justify their place in people’s baskets, especially if they’re putting prices up.

To find out whether a profitable price rise is going to be possible for your brand, you might need some help from one of marketing’s killer stats: Price elasticity, which describes how much sales volume you can expect to lose from a given price rise.

It’s very useful when deciding whether it’s going to be profitable or not.

Google’s MMM comes with bells and whistles, but not enough warnings.

MMM is hot right now. It’s because Google and Apple pulled the rug from under a competing method for evaluating online ads – multi-touch attribution – by removing 3rd party cookies.

Google’s MMM software – Meridian – has some cool add-ons compared to Meta’s Roybn:

• It can constrain outcomes to be in line with prior knowledge plus or minus a bit
• It’s granular including using smaller geographies like cities
• It incorporates data on reach and frequency for YouTube that’s not available anywhere else

Sounds clever.

The problem is that whatever modelling software you use, with MMM there are infinitely many models that will fit the data. There is no algorithm that can reliably choose which one really reflects the outside world.

Skilled and experienced modellers are needed to judge which combination of variables to collect and put in the model to accurately reflect the way that business and its marketing works.

With a lot of data science, running an algorithm is enough to get a reliable answer, but MMM isn’t like that. Throwing things at the wall, and seeing what sticks, just leads to findings that no-one believes.

The other missing warning is that Google is a media owner. One that, as customers, we trust with our searching and our email and our data, but still, a media owner with a product to sell.

It’s not surprising that Meridian offers detail on how YouTube works via reach and frequency curves, but doesn’t roll out the same red carpet for other online video that might work better for you. The emphasis is very much where Google wants it to be.

My advice is to only use it if you’re already experienced with MMM and you’re sure the bells and whistles are worth the compromise that comes from using software built by a company with a vested interest in the outcomes of your evaluation.

🚨Display has a data problem.

All those annoying ads that turn up AFTER you’ve already bought something – they’re because of an issue with data.

It’s not too little data. Programmatic display uses oodles of numbers to target people who look like they might buy something soon.

Nope. It’s an issue of interpretation.

Somewhere, some person, or, more likely, some algorithm, sees that people who get one of these ads are likely also people that buy one of these things.

But it isn’t because the ad causes the sale.

No, it’s that the ad targeted people who were going to buy anyway. And in many cases people who have already bought the blooming thing…. the example in the picture is in a lovely calming autumn palette, but it was still sort of annoying to me.

This is just one case of marketing people being misled by something that data people called “spurious correlation”.

It’s when two things – like an ad and some sales – are related, but the causality isn’t what you think.

If you want to be wise to this and a whizz with data, have a look at my Data Works course, it’s for anyone that works in marketing but hasn’t had training on marketing data.

Just because you’ve got a multi touch shaped hole in your life doesn’t mean you have to fill it with something multi-touch attribution shaped.

In fact, the advice from performance marketing legend Eric Seufert and me is that you may well be better off with a smart and experienced analyst using simple tools to reason around the data gaps than a new black box working off degraded signals.

Complete with memes, analogies, and missing each other in a small corner of London as students, this podcast is well worth a little chunk of your weekend time.

Listen to the podcast here: https://mobiledevmemo.com/podcast-econometric-advertising-measurement-with-grace-kite/

Tanks on Google’s lawn!

The most valuable real estate on the internet is the place that people go when they’re shopping. Just like the high street used to be in the olden days.

But the location is changing. The chart shows that over time, people are increasingly going to social media rather than search engines to find their information and the product they’ll purchase.

Perhaps it’s because younger people are wise to all the ways that search rankings and product reviews can be gamed, and they trust their social contacts more. Or perhaps it’s because TikTok is just so damn entertaining.

Either way, most businesses that sell online will need to take a serious look at the real estate they are renting.

Unfortunately, performance marketing decision tools are not going to be able to help with tackling this one. It’s because they look at each channel in isolation, and aren’t good at comparing search vs social or any other 2 channels in the mix. Google’s dashboards are never going to come up with a recommendation to buy more social, are they?

You can read my article all about online real-estate here: https://magicnumbers.co.uk/articles/the-two-tasks-for-online-ads/

What if people stopped searching your category?

It happens. And when it does, all of the performance marketing tools that previously seemed to control your sales disappear.

It feels like there ought to be something you can do, but there isn’t.

🛏️The chart is a case study of mattresses during COVID-19:

• The bars are an index for the number of searches in the UK. Pre-covid, there was a normal amount of demand where the dashed blue line is
• Then when the pandemic hit there was a short-lived dip in searches during the panic weeks followed by a surge in searches as people stuck at home saved on going out & invested into a comfy home
• After, in 2021, people were going out again. They had other things to do with the cash, and many now owned a fresh mattress. The number of searches was a lot lower.

When demand is high, like for mattresses during lockdown, it looks like online marketing is really working.

✔️Cost per acquisition goes down. Cost per click falls. ROAS, conversion rates and click through rates all go up.

When demand is low, like mattresses after COVID, all the indicators move in the wrong direction. Marketing suddenly looks ineffective.

But the marketing tools haven’t changed. It’s just that they have no demand to work with, and search marketing can’t generate demand. It can only harvest it.

The story in the chart was the end for more than one online furniture business.

If you still use attribution to decide how to allocate budget across channels… make 2024 the year you stop it.

It’s understandable if you do, and you wouldn’t be alone. Because once you start using it, it’s sticky.

• You’ve made decisions using it in the past, and it would be a climb-down to change them now.
• And then perhaps the figures just look good, like you’re performing really well, and your boss has got used to seeing them.
• Or maybe it’s just that the numbers are always there, and changing for something else would mean less or slower access to intel.

🤯The trouble is though, unless you are a very small and simple business, attribution isn’t good enough. The chart – by D2D and Les Binet from Adam and Eve – shows how far out it is for bigger businesses of the type that buy econometrics.

Attribution reports only half the full return on investment of brand building media channels like TV, but double the return on investment for performance marketing channels.

It isn’t easy to make the change. There will be resistance in your organisation. But it is possible.

🧪There are concrete tests you can do to find waste in your performance marketing budget and trim it. And there are good first steps that make moving into brand building safer and simpler.

My Scaling Up Works course, is hands on, practical marketing training to help you through this journey.

🐠There’s no point having a high-tech fishing net if you haven’t got a big pool full of fish.

But this is not a post about Fly Fishing by J R Hartley*. It’s about getting your marketing mix straight. Because performance marketing is the high-tech net, and brand building is what gets you a great big pool of fish.

Any advertising strategy that can help you reach a bigger audience and stay in their mind until they’re ready to buy from your category is building demand for your product now and in the future. That’s giving your performance marketing more of a chance to hook a new prospect and turn them into a customer.

📈The chart shows research on which channels are best for making your pool of demand bigger and fuller of prospects. It’s from a study of the long-term impact of 504 media campaigns. TV comes out top with online video and out of home next.

*If you don’t remember the classic Fly Fishing ad, check out https://www.thedrum.com/news/2022/06/14/world-s-best-ads-ever-96-jr-hartley-gets-people-using-the-yellow-pages-non

💼Your boss’ boss’ boss is a city analyst.

And right now, he or she probably rates brand & marketing as a very important driver of your company’s success.

To city analysts, marketing and brands are more important than leadership quality, and, sadly, a lot more important than sustainability too.

When asked why, they say it’s because raising prices is a key strategy in uncertain times, and brands are mission critical to being able to do that without losing all your customers. The last few inflation-ridden months have proved it beyond doubt.

👍It’s great news for marketing people that have to make the case for their budgets. There is an ally in the city that believes in the mission.

The trouble is, there’s a missing link. Marketing departments and city analysts are aligned, but in between them are CEOs and boards who remain sceptical.

Faced with a balance sheet that can only recognise marketing as a cost, and with profit targets to meet, they struggle to make long term decisions around marketing, to view it as an investment.

🤖Automated creativity works. AI ads outperform human ones.

But there’s a dystopia on the horizon. Not one where Skynet nuke’s the planet, but one where creativity eats itself into a boring sludge.

The chart shows brand health scores for car ads generated by AI in blue vs. people in pink. It shows that, given a set of good human ads from the past to learn from, AI ads deliver a higher brand health score, and a higher probability that you won’t air a dud.

Rationally, in the face of evidence like this, we should all switch to letting machines do our creative work. But where will it end if we do?

One possibility is that if everyone moves to AIs which are creating ads based on the same past set of good ones, all new ads will start to look the same.

The next step could be AIs that learn from each other’s past regurgitations of the original set of ideas.

🥱In a world with new ideas, potential customers will get bored, and it will all stop working.

AI may yet develop ways to avoid this awful outcome, but if not, the moral of this story is that data shouldn’t be used to dictate, instruct, or generate ideas. That job should be reserved for people.

Read more in my article here: https://magicnumbers.co.uk/articles/the-idea-is-the-bullet-data-is-the-gun/

🎙️Really enjoyed being a returning guest on ace The Sleeping Barber Podcast episode! We discussed all things upskilling your marketing team, including how to close the gap between knowing the right thing to do and actually doing. It isn’t just knowledge you need, it’s skills too.

Thanks for having me Marc Binkley and Vassilis K. Douros, loved chatting to you!

👂Listen on:

Spotify https://lnkd.in/gYGN-xfu
Apple https://lnkd.in/gMmPUwdC
YouTube https://lnkd.in/gMmPUwdC

📢If you want actual people to actually watch your ads, you can’t beat video with default sound on.

Whether it’s digital eye tracking, webcams mounted on top of TV sets, or even changes in the conductivity of people’s skin, the data shows that it’s cheaper to buy attention with TV and YouTube ads than anything else.

So, the conclusion is that your 2024 media plan ought to be MASSIVELY skewed towards video and avoid static content like display, posters, press, and Facebook.

Um, but is that right?

On the yes side: There is clear evidence that attention is related to prompted recall, brand choice, and a “virtual proxy for sales” – this is from 2 different sources, Lumen and Amplified Intelligence

But, on the maybe side: The same studies show that a little attention goes a long way. So, a little bit of attention from a different target audience on Facebook or Instagram might still be worth paying for, even if it costs more per second.

And on the side of “we need more research to be sure”: You might be surprised to hear that we haven’t found any studies yet that prove attention-optimised media plans are better for sales.

💩Is too much data a problem in marketing?

You may think that being the proud founder of an analytics company magic numbers, I’m never going to agree that data is a bad thing.

But actually, in many ways I agree with Sir John Hegarty when he says:

“Data has never created wealth, never. Creativity has, all the time. Because it imagines something… if you’re an organisation trying to increase your profitability, or expand your market, data is very important. But it’s creativity that will make that happen.”

But Brian Cox is right too. Things can only get better when he turns up.😉

It’s because creativity on its own can be misguided. An idea can look brilliant but just not turn out to be in line with how the world works.

Cox is right when he says data can assess the strength of ideas. It can test whether discoveries apply widely, or if they’re only useful in some weird corner of the universe.

🎨📊Marketing people need both things. Creativity which uniquely comes from people. And ways to test that creativity, which is the super-power of data.

*John Hegarty and Brian Cox never sat in the same room to discuss it. The conversation in the picture took place over two events at Cannes: Hegarty in 2014, Cox in 2015.