When was the last time anyone questioned the financial outcomes of retention marketing activities?

Going by the logic:

Every action you take must have a future expectation of results...

Nothing should be left to best practices or "experience".

This is the baseline of all good marketing.

And no one should ever settle for less.

Yet this is pretty typical for many agencies...

Trust-me-I'm-an-expert "strategy".

Nurturing customers, meeting the customer where they're at in the journey, personalisation.

Makes sense in theory, but what’s the incremental lift?

Is it the most efficient use of your limited resources?

At the 7-figure mark, trying to break 8-figures, this is an ever more crucial time.

That's what CTC's Global Accelerator Program solves for.

1. Better alignment with traditionally siloed growth <> retention teams.
2. Better agility to support a very dynamic marketing calendar.
3. Better analysis of activities that impact your P&L.

Together with Growth Strategists, Retention Strategists make this work.

At CTC, we always sign our names on the bottom line.

Because we know that’s what it really means to win.

If this speaks to you, let's chat and see if we'll be a good fit for each other.

Josh Tay

Some good stuff from this month.

Because everyone still asks to see designs. So here we go :)

Director of Retention, Common Thread Collective | Building the world’s strongest retention program for 7 & 8 figure DTC brands

Just because you sell skincare, doesn't mean you need to only talk about skincare.

In this example, we tested running traffic to 2 quizzes.

Quiz 1: Which smoothie matches your personality
Quiz 2: Which beauty routine is right for you

We didn't expect the smoothie version to outperform by so much.

Most people would've stopped just after creating a quiz.

Or maybe compared quiz AOV/LTV vs non-quiz takers.

But we thought this would be a worth-while test.

Lesson: Think about your customers' lifestyle. 

And alternatives (not complementry) for your product. 

Weave them together.

Now this is a fascinating insight that our retention team passed to the growth team.

Client industry: Apparel

And we found that customers who bought designs with dogs had much higher LTV than those who didn't.

What brand will be doing:

- Add new filter for dogs for easier nav
- Test new designs with dogs

What we'll be doing:

- Dedicated dog ads
- Find more of such niches

How much time is your team investing into drawing insights?
Is your team reporting on metrics or taking action?
Are you asking the right questions?

Giving discounts in your flows to increase AOV?

What criteria do you have to determine who receives a discount?

Everyone?
Non-customers only?
Discount with threshold? But which amount?

To answer that, you'll need to look at your AOV histogram.

1. If you based it off AVERAGE order value, you might be pretty off

- Averages lie
- If we had set it at the mean, there'd be a very big chunk of people who wouldn't have moved
- Purely because their cart value might've been too far (see how far mode is from mean)
- Mean is getting pulled up by select customers who spend enornmous amounts
- Question is, where are most of your customers?

2. Look at MODAL order value for help

- Your objective is to move people one bucket up
- Moving the bulk of customers up will translate to higher AOV by default
- It's the same for how you'd create free shipping thresholds or free gift with thresholds in your gamified cart drawers.

There's math to most of these things.

Only do things when it makes sense for the business.

You might win on revenue, but at what cost?

(Long post ahead!)

Today I want to breakdown a real post-mortem of a recent play we did for women’s headwear brand.

And exactly how we still hit the forecast while mitigating downside risk. 

Context:
Client wants to repeat last year’s promo mechanics and marketing calendar.

Expectation:
Client looking for 1.5x rev growth with 20% higher CM.

So unfortunately, that means getting more out of email and sms channels

Problem:
Based on projections following last year’s cadence, we weren’t likely to hit target revenue desired by client.

Constraint:
At the end of the day, it’s the client’s business, and we’re here to deliver results. So it’s our job to best solve for it.

Is this sustainable? 

Nope. But it’s the best stop-gap measure before we have a conversation on expectations, while still delivering on performance to prove we can.

👀 Observation of historical:

Every major promo (BFCM, Mother's Day, Labor Day, Independence Day) did generate more revenue YoY. 

Revenue up 87%, but they had to send to 133% more people to get there. 

At some point, we’ve just been adding subscribers who don't convert into the campaign mix…

Dragging down the economics of the entire channel.


🧠 Insight 1:
RPR declined every single time—ranging from -7% to -42%. 

What was happening is that they were running the same plays to bigger lists and getting diminishing returns per subscriber.

It was reusing the same promo mechanics promo on promo, because “it worked”.

And expanding audiences toe extract more revenue.

What was “working” was going to crash real soon.


🧠 Insight 2:

The correlation between campaign frequency and RPR was actually positive (+0.25).

More campaigns in a month didn’t tank performance. 

What does correlate with declining performance though, was list size growth:

Early period: 56K avg recipients/campaign → $0.0326 RPR
Late period: 126K avg recipients/campaign → $0.0261 RPR

As the send sizes expanded 2-3x, engagement metrics collapsed.


🔬Analysis:

Most would think we have a list fatigue problem.

Initial thought is to reduce sends.

But we actually don’t have a demand elasticity problem.

We are just dealing with the law of shitty cohorts.


👎 The wrong things to do:

1. Expand audiences to get more impressions.
(This will trigger law of shitty cohorts)

2. Reduce send volume
(This wouldn’t allow us to hit rev targets)


👍 What to do:

1. Increase send volume to tighter engagement segments since list has been proven to be able to handle.

2. Product promos only to go out to product-interest segments.

This is why we never skip forecasts.

The forecast allows us to gap-find and it’s based on the exercise of rigorous forecasting that we uncover these insights that we’d otherwise have missed out on, which might have resulted in incorrect execution.

When creating any post-purchase upsells/cross-sells, always rely on your product journey map.

The common way:

Person adds sample > They buy regular next

Data might show otherwise.

In this case, person goes straight into bundle.

Even for singles, they seem to not replenish the same product.

Instead, they gravitate towards bundles, and stay on bundle orders.

This could drastically change the way you push for repurchases.

Interestingly, this mystery item was the fourth most purchased item.

This item's added to the cart drawer as a default upsell on every product.

Avg item price: $35
Mystery item: $15

Objective: How can we liquidate slow/dead inventory faster?

Go-forward questions:

1. Should we create a mystery bundle option too?
2. Can we create "mystery gift" as a front-end offer?

These will all be tests to be conducted.

Ask the right questions...

... Come up with the right hypotheses

... Test the right things

... Take the right actions


Client niche: Accessories

Klaviyo audits alone are insufficient to determine what to do for your retention program.

How should you prioritise activities?

Of the flows recommended, what will the content/angle be?

It's one of the reasons why we look through everything.

Based on this finding, we're:

- Doubling down on resources for refer-and-earn program
- Creating multi-bundles and quantity break discounts
- Include more creative for influencers in pre-purchase flows

(we found that most people who said instagram, youtube, tiktok, also meant influencers and creators)

To do a sanity check, we looked at codes redeemed and saw that most first time customer codes redeemed were from affiliates.

Sometimes things like these we're overthinking.

But we're actually seeing higher lead:conversion efficiency across the board in the last 90 days vs previous period.

So we know that's working.

---
Client niche: Fitness equipment

4 mistakes I made when first forecasting email revenue and how you can avoid it too.

I wish we could figure this out faster too.

But the retention scape in this aspect is still in its infancy.

We're doing our best to fail faster.

To come to a clearer truth.

Here's 4 things I got wrong and how we're solving for it now:

1. Not taking monthly volume into account (rev degradation)

❌  As we increase campaign volume, unsubs might increase and RPR might decrease.

What should've been done was to create a weighted average of RPR based on volume of sends per week over time, split by promo vs non-promo

---

2. Not taking post-promo revenue dips into account

❌ From the last promo sent, you need to determine the "cooldown period" before demand picks back up again.
---

3. Not taking last-year-same-period's email volume into account

❌ For the sake of simplicity, I naively assumed last year's RPR to be same as this year's. 

I just applied a seasonality factor on top with a YoY growth factor.

But the diminishing returns from higher volume played a bigger role than I anticipated.
---

4. Not discounting active subscriptions in email campaigns

❌ If you send a general educational campaign to your engaged list with active subscribers in there...

You'd have been inaccurately attributing revenue.
---

Now we're factoring all that and building it into our forecasting algorithm.

And fully automating this to eliminate human error.

Humans are terrible at extrapolating data.

Worse if we need to incorporate several other aspects.

Tech enablement is the future of marketing.

Built upon solid reasoning, 
asking the right questions, 
with an endless desire to improve, 
and relentlessness to follow-through on the outcomes we promise on

Your cart drawers, checkout upsell, post-purchase upsell funnel will bump your LTV up more than a bounce-back post-purchase email will.

On these pages, you have 100% open rates.

With the highest purchase intent.

Take advantage of it.

Different products in cart might need different upsells based on:

- LTV growth by product
- Total LTV by product (including AOV)
- Balance with product demand (volume)
- Product journey
- Frequently bought together

Then once you have your various funnels/upsells in place, there should be monthly measurement to identify what needs improvement.

How equipped are you and/or your retention team right now to execute this?

Reviews immensely help reduce friction to the purchase.

People look at the number of reviews to know if this brand/product is trustworthy.

It makes sense to build your review war chest.

In post-purchase surveys, we've asked specifically about reviews.

Question 1: "What made you choose to shop with us?"

Options: "Reviews", "Referrals from friends", "Video tutorials", "Others"

Of those who selected Reviews, we had a logic to ask the next question...

Question 2: "How did the reviews help you make a decision today?"

Options: "Other customers' stories", "Photos/Videos of products in use", "Number of reviews vs other brands"

And we saw high photo/video review responses.

This isn't a theory propogated by review apps.

We personally tested this.

That's how we know what to do.
(Always be skeptical and own your own data. Everyone on the internet lies. Maybe I'm lying too)

But if you're too lazy to test, here's what you can do to set up.

1. Go to your review app
2. Create coupon tiers
3. Customise review request emails to let customers know about the added incentive
4. Set an expiry on the discount
5. Send reminders to customers who left respective reviews, to claim their discount

Optional:
A/B test with Intelligems or a similar tool, with and without reviews on your PDP.

Or with the number of reviews stars right above your product name.

And see how that affects conversion rate.

Who said retention was only emails and sms?

Is your retention person supporting you with these?

Always nice to know we're pacing on target, within <10% delta.

We still have big moments we're expecting towards end of month.

Tracking performance this way takes away guesswork.

The worst thing an agency can do is to come up with excuses to justify underperformance.

Over and over again.

When all that was needed was accountability.

We're now more than 90 days past BFCM.

You know what that means?

Time to reactivate your black friday shoppers who are unengaged on digital.

Looking at your repurchase windows, you'll see that majority of LTV lift for new customers happen within first 90 days. 

(typically)

If you're a brand with replenishable products...

Or you sell things with high novelty/"new drop" factors...

This is definitely something you'll want to get in on.

Relook at past year's data and you'll also find this:

Your black friday buyers are a lot more sensitive than other promo period shoppers/evergreen.

(I've had a post with some of this data before if you wanna check that out.)

So if we're nearing the end of the highest probablity curve of repurchase and they're still not coming back...

Much less engaging...

You'll want to do something about that.

1. Offer
Again, we know holiday shoppers typically only buy with discounts

2. New channel
If they're not engaging on email, sms, try new mode of communication

And this is the perfect time to start.

+ you only have a very short window to execute.

You'll want to run your winbacks for this group of customers before the next major sale.

Which for most US brands, is going to be Memorial Day and 4th July Sales.

If you start right now, you'll need a couple of days to get post cards delivered to doorsteps.

Then, people don't take action immediately.

Postcards might sit on their desks or coffee tables for a while.

That's why we wait out with a 60-day maturity period.

🚨 IMPORTANT:
You don't want your winback offer clashing with your site offer.

Because now that just muddies data and the offers cannibalise each other.

Timeline:
Now > 14 Mar = Plan your campaign, offers, creative, segments
15 Mar > 20-ish Mar = Delivery
20 Mar > 20 May = 60-Day Maturity Period

Just in time for memorial day promo launch date.

And you can do your usual holdouts.

But this is a crucial time to strike.

If you want to run a proper winback campaign, do it once, do it right.

Have an Amazon store that gets more traction than your .com?

Consider an exit intent popup redirecting traffic there.

Targeting:
- triggered on PDPs only
- $0 cart value
- stayed more than 15sec
- or scrolled more than 50%

Create specific popups for each key product, with relevant redirects.

Instead of wasting that visit, send them to somewhere you know people are actively converting on.

We're doing this with a home improvement brand that sells DIY light kits right now.

If people want to shop on amazon, we can't force them not to.

Return rates on amazon for this brand aren't crazy high.

So this will be our interim solution before the growth team solves for the conversion issue.

There's no one way to use popups.

Everything is just a tool.

What matters is how you use it to benefit your business.

“Bundles increase AOV”

Logically, that should translate to higher LTV.

So many people default to bundles, which might not always work.

In this case, I’m specifically talking about pre-set bundles, which is the most common way of bundling. 

Sometimes customers just want to pick out exactly what they want.

Forcing a bundle may actually end up reducing conversions.

Ideas like:
Hero product with slow movers
Hero with high LTV product to push profit
Hero with accessories that go well together

These are valid pairings.

But again, we’re assuming people want the next product we’re recommending.

What you want to sell might not be what customers want to buy.

There are two ways to go about this.

Product journey + frequently bought together analyses

Which I’ve talked about before in previous posts.

Let people choose on their own

Aka Build Your Own Bundle.

This works especially well (and much easier to implement) for brands with similar AOV products in the catalogue.

And you can gamify this with the typical tiered discounts with threshold.

I’ve seen bundle take up rate go up to 20% with dedicated BYOB pages in the nav bar.

Example 1 
We have a client selling canned tea. They have 12 flavours. Instead of forcing 6 packs of single varieties then force customers to buy multiple carton to try different flavours, you select your own quantity of flavours for each 6-pack.

Structured like a quiz

Step 1: Select how many cartons you want
Step 2: Carton 1 - Which flavours do you want
Step 3: Repeat for however many cartons selected
Step 4: Would you like to add another carton for X% off? (Progress bar of discount at top of page to)

If AOV varies widely, adopt a similar strategy as how you’d buy a computer.

Base product
+ one add on from category 1 at $X
+ one add on from category 2 at $Y
+ one add on from category 3 at $Z

Each of these are optional.

Example 2
Another client selling sofas. That’s base product. Category 1 are removable spill proof covers. Category 2 are pillow sets. Category 3 is a 2-year warranty.

In our frequently bought together widget below, we still have opportunity to cross-sell sofa extenders and poufs.

Now you give customers lots more flexibility while maintaining some level of control of options.

No fluffy customer journey theory.

Just real stuff that you can implement right now.

And all the data points you need.

Feel free to take this and run with it.

Client niche: F&B, subscription

How to overcome weak subscription program take-up rate

"Multi-month" cohorts almost always will have higher LTV than monthly subscription cohorts.

More specifically, I’m talking about cumulative LTV.

Which is at the end the most important thing to the business. 

That $ value.

So this is definitely something every subscription brand should be testing.

In this example, brand sells soaps.

- One-time orders
- 15% monthly subscribe and save
- 19% for double pack bundle
- 23% for triple pack bundle

Your CAC might be higher on these multi-packs.

Largely also due to much higher AOV.

But not only is cumulative LTV higher, margin LTV per customer beats monthly subs hands down.

Essentially you’ve collected ltv upfront. 

Which is great. 

You don’t have to worry about churn as much. 
CAC payback’s shorter. 
So is your cash conversion cycle.

Definitely a test worth trying.

LTV comes in many forms.

Do what's right for your financial outcome.

Are you levergaging Web Push notifications?

- Welcome
- Browse abandons
- Cart abandons

If you find your lead:conversion rate on these email/sms flows are weak, it might make sense to have another channel to support.

At the end of the day, it's about impressions.

So where do your customers hang out, that you also be on?

Show up there and drive them back to site.

Then track lead:conversion efficiency

The way you structure a tiered loyalty program must take psychological triggers in mind + factor in unit economics.

Weak participation rate = Poor incremental revenue

= unnecessary complexity 

= useless extra cost to run the program

Three post-purchase micro surveys conducted across three different brands.

All had one same option:

"When there's a sale on"

In all instances, that came up as the highest volume response.

This is probably the reality for most ecommerce brands.

People are most likely to shop when there's a sale or a discount.

Knowing this, it's ever more important to bake the discounts already into the price of your product.

So that you can afford to "discount" when necessary.

Without having a really strong, notable brand, is it really possible to never go on sale, and still maximise your growth potential both on new and returning rev?

What do you think?

Scaling eCommerce Brands Through Profitable Acquisition

Driving profitable, scalable growth for 8 and 9-figure eCommerce brands through integrated performance marketing, creative, and data systems.

AFR Fast Starter 2024 | Finalist Melbourne Young Entrepreneur | Startup of The Year | Emerging Agency of The Year

At Blue Sense Digital, we work with 8 and 9- figure eCommerce brands to align growth strategy with financial outcomes and operational leverage.

We work with businesses where marketing, finance, and operations need to be aligned for profitable scale.

Our clients typically operate in:
· Fashion & Lifestyle
· Consumer Goods
· Home & Living

We believe profitable scale depends on three critical levers:

→ Creative Velocity & Diversity
→ Technical Account Structure
→ Data Integrity

By integrating these pillars, we help brands build a scalable acquisition model, tied directly to financial outcomes.

For earlier-stage brands, we share proven frameworks and strategies through educational content and resources:

→ https://www.youtube.com/@bluesensedigital

If you’re an eCommerce leader ready to build your next chapter of growth, let’s connect: 
https://calendly.com/bluesensedigital/exploratory

I promise to honour your time.

Nathan Perdriau

We are looking for A-players who are seeking rapid skill development (outside of the ad account).

You will:
→ Be implementing across 8-9 figure retailers 
→ Step into a clear roadmap towards leadership from day one 
→ Learn from our lead strategist, myself & team leaders to accelerate your development

You will learn:
→ Financial analysis & forecasting 
→ DTC & omni-channel growth strategy
→ Data science for causal experimentation
→ Creative strategy & planning

If you have 3+ years of experience in performance marketing, then I'd love to talk. 

It DOES NOT matter where you're based globally. We hire from everywhere.

Please do NOT message me if:
→ You don't have incredibly high standards for yourself
→ You don't take accountability for everything around you
→ You aren't a perpetual student

If you think you're a fit, send me a message.

And if you know a sharp, experienced marketer who's outgrown their current role - please pass this along.

Thanks.

Producing Better Financial Outcomes for 8-9 Figure Brands

Ads in 2026 are unforgiving.

Miss one of these and you’re playing on hard mode.

Here’s what actually matters right now.

1. Your base Meta structure
   Manual campaigns are basically gone.
   Everything is Advantage+ now, with manual control layered in.

The simplest version that works:

- One testing campaign with low budgets
- Winners get moved via post ID into a scaling campaign

Not perfect for every account.
But a solid default.

2. AOV and LTV are non-negotiable
   Trying to make Meta work on a $30–$40 product?
   In the US?
   Good luck.

Reality:

- Sub $75 AOV struggles
- Sub $100 AOV is brutal unless cohorts carry it

CPMs don’t care about your margins.

3. Partnership ads still matter
   Not a “hack” anymore.
   Too saturated for that.

But still one of the few levers that moves results.
Especially founder-led and creator-led pages.

And yes, they behave differently to standard ads.
The page data matters.

4. Creative investment wins accounts
   This is the big one.

If you’re spending $300k/month on Meta and:
→ $10–20k on creative

You’re underinvesting.

If your competitor spends:
→ $300k on media
→ $60k on content

They will beat you.
Even with worse buying.

In 2025, creative is the strategy.
Media buying just distributes it.

Budget accordingly.

undefined undefined

Comment

Watch our guide here: https://www.youtube.com/watch?v=DnpYApVNu-Y

teagan youtube

blue-sense-digital

blues-brothers-podcast

Emmanuel Daniel

Every time you launch an ad set, all the spend goes into just one ad.

How do we fix this?

How do we get some spend into ads 2 and 3?

We spent all this money and time making these creatives. We want to know if they worked.

Four options:

Option 1: Relaunch without ad 1

Force spend through the other 3 ads.

Is that a good idea?

It'll let you test these ads, but there's probably a reason they weren't getting spend.

They're not going to perform as well as ad 1.

If you're maximizing efficiency, not a great move.

Option 2: Do nothing

Just leave these 4 ads running.

2 won't get any spend. Oh well.

Meta is really good at distributing spend, so we trust it.

Option 3: Set up creative testing

This is a new feature Meta rolled out about 2 months ago.

You select the ads you want even spend distribution through and it will distribute spend accordingly.

Option 4: Make new creatives

Make more creatives like the one getting all the spend.

What's the best option?

Best: Make new creatives.

Make more creatives like the one getting all the spend.

Second best: Do nothing.

There's a reason these ads aren't getting spend. They probably won't perform well.

Is that true all the time? No. 5% of the time, Meta might be wrong and you need spend through them.

But 95% of the time, you'll be wrong trying to distribute spend here.

Third best: Set up the creative testing feature.

Worst: Duplicate and turn off the high-performing ad.

I've never really seen that work as a strategy.

The lesson:

Stop trying to force spend into underperforming ads.

Either trust Meta's distribution or make more winners.

Don't fight the algorithm just to test ads that probably won't work.

Make more of what's already working.

That's how you scale.

If you’re running Advantage+ and judging ads by ROAS alone…

You’re probably scaling retargeting.

In a post-segmented world, we don’t structure accounts as:

* Top of funnel
* Middle of funnel
* Bottom of funnel

Everything sits inside one broad campaign.

So you’ve lost visibility.

Here’s the metric that replaces funnel labels:

Frequency.

How to use it properly:

Step 1
Look at the last 7 days only.
Not 30 days. Not lifetime.

Step 2
Add frequency at ad level.

Step 3
Interpret exposure depth:

* 1–2 = mostly cold
* 3–4 = warming
* 5+ in 3–4 days = heavy retargeting

Real example from a large spend account:

Ad 1
Frequency: 7 (in 3 days)
Purchases: high
ROAS: strong

The team’s instinct?
“Scale it. Make more like this.”

Wrong.

It’s winning because it’s serving warm / bottom-of-funnel users repeatedly.

Now compare:

Ad 2
Frequency: 1.8
Purchases: lower
ROAS: weaker

Most teams pause this.

But this is likely the best true cold-acquisition creative in the account.

And cold performance is what determines the ceiling.

If you ignore frequency, you’ll:

* Overproduce retargeting-style creative
* Underinvest in prospecting angles
* Hit a growth plateau you can’t explain

Here’s the system I recommend for $5m–$500m brands:

1. Categorise ads weekly by frequency band.
2. Judge performance inside its exposure bucket.
3. Scale cold winners separately from warm converters.
4. Build creativity intentionally for each depth of awareness.

Frequency isn’t just a delivery metric.

It’s your funnel diagnostic in consolidated accounts.

Most brands optimise for visible ROAS.

Operators optimise for where the conversion actually originated.

What do I actually do day-to-day at the agency?

The highest leverage opportunity within an agency is people.

An agency is fundamentally a recruitment and training business in that order.

Then sales and marketing.

Here's exactly how my time is split:

40% training

All the ancillary activities associated with developing the team.

30-40% client delivery

Strategy, quarterly planning, account oversight.

20% sales-related activities

Audits, content, ads.

(I batch everything so it's very time-efficient.)

My co-founder spends 40% of his time on recruiting alone.

The key takeaway for agency owners:

If your time is disproportionately spent on sales and marketing, you're probably not running a very good agency.

You shouldn't be demand-constrained.

You should have infinite demand because you should actually be good.

The real constraint on a good agency is hiring and talent.

That's where a lot of your time should go.

And here's the key for eComm brands vetting agencies:

Ask the founders directly:

"How much of your time is spent on recruiting, hiring, and training?"

Because that is what makes an agency good.

If time isn't being spent there by the founders, it's probably being spent there by no one.

And therefore, it's just a sales and marketing machine focused on acquiring clients.

Not on actually serving them well.

What to ask when vetting an agency:

1. How do you recruit your team?
2. What does your training process look like?
3. How much time do the founders spend on training vs sales?
4. What's your team retention rate?

The best agencies are talent obsessed.

Not lead generation obsessed.

If they can't answer those questions clearly, keep looking.

This Meta setting could be hurting your performance.

It’s called "maximise number of conversions.”
It’s the default. You’re probably using it.

The problem is simple: it optimises for the lowest CPA, not profit.

That sounds good at first. Cheaper purchases should mean more sales, right?
Not if you sell multiple products.

Example:

Product 1: Hoodie ($90)
Product 2: T-shirt ($50)

Costs:
Hoodie delivery: $21
T-shirt delivery: $13

CPA:
Hoodie: $45
T-shirt: $25

Both products are at 2x ROAS.
T-shirt: spend $25 to make $50
Hoodie: spend $45 to make $90

Same ROAS.

But Meta will push spend to the t-shirt because it has the lower CPA.
Meta doesn’t care about ROAS or profit. It cares about CPA.

Now look at profit contribution:

T-shirt: $50 - $13 - $25 = $12
Hoodie: $90 - $21 - $45 = $24

The hoodie is twice as profitable.

But if you’re using “maximise number of conversions,” Meta will still send most of the budget to the t-shirt.

This shows up a lot in fashion accounts.
A low-priced item gets added to the campaign and suddenly eats all the spend.

Why?
Lower price = lower CPA = algorithm favourite.

The fix:

If you want higher AOV, better profit contribution, and smarter budget allocation across products, switch to “maximise conversion value.”

That tells Meta to prioritise value, not just cheap purchases.

Stop optimising for the cheapest conversion.
Start optimising for the most profitable one.

If you still think Meta targeting = interests & look-a-likes…

You’re 5 years behind.

Here’s how targeting actually works in 2026.

Back in 2019:

* Someone engaged with pet content
* Meta slapped a “pets” label on them
* You targeted the “pets” interest
* Ads were shown to everyone with that tag

Simple.

But, people being “interested in pets” tells you almost nothing about:

* Income
* Location
* Buying behaviour
* Psychographics
* Brand affinity

So Meta changed the algorithm.

They moved from label-based targeting
→ to vector-based targeting.

Here’s what that means in plain English:

Instead of tagging someone as “pets”…

Meta positions every user inside a multi-dimensional space.

Every:

* Post
* Reel
* Story
* Click
* Purchase

Pushes that user in a certain direction.

Over time, people naturally cluster.

Example:

* High-income mums who like pets
* Urban males into outdoors + premium brands
* Fitness-focused students in major cities

They don’t share one label.

They share a pattern.

Now here’s the important part for brands:

When you run broad…

Meta doesn’t “ignore targeting”.

It tests your ad across the space.

Then it identifies which cluster responds.

Then it doubles down on users who sit near that cluster.

Your creative becomes the targeting filter.

That’s why:

* Weak creative fails in broad
* Strong creative scales aggressively

And why layering 20 interests does almost nothing now.

Practical takeaway:

If you’re running ASC or broad campaigns:

1. Stop obsessing over interest stacks.
2. Focus on clear ICP-aligned messaging.
3. Analyse who is actually buying (geo, AOV, device, age bands).
4. Feed that insight back into creative angles.

Your job isn’t to find the audience.

It’s to create ads that attract the right cluster.

Back with ABOs and CBOs, here's how they worked:

They'd serve to 1,000 random people.

Once you accumulated 50 conversions, Meta knew who to target.

Then it would exit the learning phase.

But Meta realized this is a huge waste of spend.

So instead, with Advantage+ campaigns, it actually skips the learning phase by using historical conversion data within the account to go after the highest-intent people possible.

That's why when you launch Advantage+ campaigns, performance starts really good and then dips off.

It capitalizes on all the high-intent audiences from the historical conversion data and targets those people first.

Whereas with ABOs, you'd typically see better performance over time because they'd start collecting conversion data and then learning.

The Advantage+ lifecycle:

Day 1-3: Great performance (hitting historical high-intent audiences)
Day 4-7: Performance dips (exhausted high-intent audiences)
Day 8+: Performance stabilizes (finding new audiences)

This is normal.

Don't panic and turn off campaigns when performance dips after the first few days.

It's just moving from high-intent to cold audiences.

The fix:

1. Expect the dip
2. Give campaigns 7-14 days to stabilize
3. Feed them with more creative to expand reach
4. Scale once stabilized, not during the high-intent phase

The lesson:

Advantage+ skips learning phase by using your account's historical data.

This is good (faster to profitability).

But it also means you'll see a performance dip as it exhausts high-intent audiences.

This is expected behavior.

Don't kill campaigns too early.

Let them stabilize.

The Meta “learning phase” won’t save your bad ads.

If you’re at 1x ROAS today…

You’re not magically hitting 3x because the status changes to “Active”.

This belief is burning cash in $5M–$500M Ecom accounts.

Here’s what founders get wrong:

They launch a new ad set.

Days 1–5:

* ROAS = 1.0
* Break-even = 3.0

Response?

“Let’s just wait until it exits learning.”

That’s not how it works.

What the learning phase actually does:

→ Reduces volatility
→ Improves delivery stability
→ Tightens audience matching

It does NOT:

→ Double your ROAS
→ Fix weak creative
→ Save bad economics

Before learning completes, results look like this:

Up.
Down.
Spike.
Drop.

After learning completes:

More consistency.

Less variance.

That’s it.

In most accounts, performance lift post-learning is:

* 0–10% typical
* 10–20% good
* 30% rare

If you need a 200% improvement to be profitable…

You don’t have a learning phase problem.

You have:

1. A creative problem
2. An offer problem
3. A margin problem
4. Or a targeting strategy issue

Here’s how to approach it properly:

1. Judge direction, not perfection, in first 72 hours
2. Compare CPA to your allowable, not just ROAS
3. If you’re 60–70% of target, it may stabilise in the range
4. If you’re 30% of target, kill it

Stop waiting for the algorithm to rescue you.

Meta optimises.

It doesn’t perform miracles.

If the foundations aren’t there…

The learning phase just stabilises losing money faster.

Most brands guess how many ads to launch each month.

That’s why creative output never matches revenue goals.

Here’s the equation you actually need:

Winning Ad Value × Hit Rate = Expected Value per Ad Launched

Break it down.

1. Winning Ad Value
   How much revenue does one true “winner” generate per month?

Not a decent ad.

A real scaler.

2. Hit Rate
   Out of every 10 ads you launch…
   How many become winners?

If 1 in 10 wins, your hit rate is 10%.

Example:

* A winning ad drives $100k/month.
* Your hit rate is 10%.

That means:

Each ad you launch is worth $10k in expected monthly revenue.

Now reverse it.

If you want to add $300k/month in new customer revenue…

You need to launch 30 ads.

Not 5.
Not “as many as we can manage.”

30.

This is how you remove emotion from creative strategy.

Here’s how to apply it this week:

1. Pull 6–12 months of Meta data.
2. Define what qualifies as a “winner”.
   (Spend threshold + CPA target.)
3. Calculate your real hit rate.
4. Work out average monthly revenue per winner.
5. Set a revenue target.
6. Reverse engineer required ad volume.

Then do one more thing most brands skip:

Back-propagate into diversity.

If you need 30 ads per month and your hit rate improves when you test:

* 3 hooks
* 3 angles
* 2 formats
* 2 audiences

You can plan that in advance.

Creative stops being random.

It becomes maths.

$5M brands ask:
“What should we test next?”

$50M brands ask:
“How many tests do we need to hit the number?”

Volume isn’t chaos.

It’s controlled probability.

Know your numbers.

Then scale your output to match your ambition.

If one ad is getting all the spend…

Stop trying to “fix” it.

Every eComm team has done this:

You launch 4 ads in one ad set.

Ad 1 gets 90% of the spend.

Ad 2, 3 and 4 get scraps.

And someone says:

“How do we force spend into the others?”

Here are your 4 options — ranked.

1. Make more creatives like the winner.
   Best option.

If Meta is pushing spend into one ad, it’s because it’s predicting higher return.

Instead of fighting it:

* Break down the winning ad.
* Identify the hook.
* Identify the angle.
* Identify the format.
* Recreate variations.

Scale what’s working.

2. Do nothing.
   Second best.

Meta’s distribution isn’t random.

95% of the time, if an ad isn’t getting spend, it’s because it won’t perform.

Trust the signal.

Optimizing for ego (“we spent time on that creative”) kills efficiency.

3. Use Creative Testing.
   Third.

Meta rolled this out recently.

You can select specific ads and enforce more even spend distribution.

Useful if:

* You’re validating messaging.
* You’re early in account learning.
* You need directional data.

But understand:

You are overriding the algorithm.

4. Duplicate the ad set and remove the winner.
   Worst option.

Yes, you’ll force spend through the others.

No, it won’t magically make them good.

If they were better, Meta would already be spending on them.

Here’s the mindset shift:

You’re not in the “fair distribution” business.

You’re in the profit business.

For brands doing $5M–$500M per year:

Creative volume > creative control.

When one ad wins:

* Don’t redistribute spend.
* Don’t overanalyze the losers.
* Double down on the signal.

Most scale comes from pressing advantages.

Not rescuing underperformers.

Every time you launch an ad set, all the spend goes into one ad.

But you spent all this time making these other ads.

How do we get spend into the others?

Four options:

1. Relaunch the ad set

Duplicate it and hope the next time spend goes into the other ads.

2. Do nothing

Just let it distribute how it's distributing.

Meta is probably better than you at knowing where to put spend.

3. Use the creative testing feature in Meta

Select your ads and set up a creative test that will force spend through the other ads.

4. Make new creative

What should you actually do?

Best option: Make new creative.

There's a reason Meta is distributing spend like this.

It's probably because this is the best-performing ad at this level of spend.

Second best: Do nothing.

Let Meta optimize.

Third option: Use the creative testing feature.

Last option (don't do this): Relaunch and duplicate.

It's going to do absolutely nothing.

It's a waste of time.

Why this matters:

Meta's algorithm is very good at finding the best-performing ads.

If all the spend is going to one ad, it's because that ad is winning.

Trying to force spend into underperforming ads won't help.

What to do instead:

1. Accept that one ad is winning
2. Analyze why (hook? messaging? format? avatar?)
3. Make new creative inspired by what's working
4. Launch new ads and let Meta test them

If the new ads are better, Meta will shift spend to them.

If they're not, Meta will keep spending on the winner.

Stop fighting the algorithm.

Work with it.

Make better creative.

That's the only lever that actually moves the needle.

Should you run interest or broad targeting in 2026?

It comes down to how machine learning inside Meta Ads actually works.

How Meta used to work:

They’d collect data and label you.

Engage with pet content?
You get tagged with the “pets” interest.

Simple. Keyword-based buckets.

That’s largely outdated.

What it moved to: vector-based targeting.

Sounds complex. It’s not.

Instead of assigning a word like “pets,” Meta assigns you a position in multi-dimensional space.

Think of it like coordinates.

If you engage with pet content, you get pushed slightly in that direction.
Engage again? You move further.

But this isn’t 3D space.

It’s hundreds or thousands of dimensions.

Every interaction pushes you slightly in different directions based on:

* Content type
* Frequency
* Recency
* Purchase behavior
* Demographics
* Device usage
* And countless other signals

Over time, people naturally cluster.

Groups form:

* Parents cluster together
* Young males cluster together
* High-income hobbyists cluster together

Everyone sits in a unique position, but similar people sit near each other.

So what’s the difference?

Interest targeting relies on old-school labels.
“Target people interested in pets.”

Broad targeting leverages this entire vector space.

When you launch broad, Meta shows ads across many clusters.
Then it quickly narrows into the pockets where purchases actually happen.

It’s not just asking:

“Do they like pets?”

It’s asking:

* How much do they like pets?
* What else do they like?
* What behaviors correlate with buying?
* What demographic + economic signals compound that?

Maybe your buyers aren’t just “pet lovers.”

Maybe they’re:
Pet lovers

* Aged 50–55
* 4 kids
* Top 10% household income
* Recently browsing eComm sites

Interest targeting can’t capture that nuance.

Broad can.

Why broad wins in 2026:

1. It finds buyers you’d never manually select
2. It models nuance, not just surface-level labels
3. It adapts in real time based on conversion data
4. It compounds signal density over time

Interest targeting = static labels.
Broad targeting = dynamic, multi-dimensional modeling.

In 2026, machine learning is better at finding your customer than you are.

Run broad.

Let the algorithm do its job.

How many creatives do you actually need per week to scale?

There's a good way to solve this:

1. Figure out your hit rate (your average rate of getting a high performer)
2. Figure out how much spend a high performer can hold on average
3. Then you've got your maths

Example:

You want to spend $10K/day.

You know your hit rate is X%.

You know each high performer can hold Y amount in spend.

You just back-propagate to the amount of ads you need.

If you can't be bothered doing that, here's an easy back-of-napkin number:

For every $1,000/day in ad spend, one new ad is generally going to get you in a really good position.

The formula:

Daily ad spend ÷ $1,000 = Number of new ads per day

Examples:

$10K/day → 10 new ads/day → 70 ads/week
$50K/day → 50 new ads/day → 350 ads/week
$100K/day → 100 new ads/day → 700 ads/week

This is a rough guideline.

Your actual number will depend on:
- Your creative hit rate (how many ads become winners)
- Your average spend per ad (how much each ad can hold)
- Your niche (some niches need more diversity)

But as a starting point, 1 ad per $1K/day is solid.

The lesson:

Most brands don't launch enough creative.

They're trying to scale to $10K/day with 5 new ads per week.

That's not enough.

You need 70 ads/week to sustain $10K/day.

If you're not launching enough creative, you'll hit a ceiling.

Scale your creative production with your ad spend.

Otherwise, you'll fatigue your ads faster than you can replace them.

The issue with CBOs when you don't have a lot of conversion data:

They over-prioritise soft metrics.

In fact, this is how the entire Meta algorithm works.

When you first launch a campaign or ad set, it's not actually optimising for purchases.

Because there are none. Or at least aren't a lot.

So instead, where there's a lot more signal in clicks and impressions.

Meta is going to prioritise all of these soft metrics - to decide how to distribute spend.

That's why if you're running an account with a very high average order value (so you don't have a lot of purchase volume turning through), and you run an ABO structure...

What ends up happening:

You might have 3 ad sets and all the spend just ends up going into the top ad set.

Not because it's driving the best financial outcome for you.

But because it simply has the highest CTR and the lowest CPC.

You have to be really careful about this, particularly when structuring the account.

The lesson:

Soft metrics get prioritised, at least within the first 1-7 days of something new getting launched.

This occurs under ad sets and under campaigns.

Be careful when:
- Running high AOV products (low conversion volume)
- Mixing images and videos in the same campaign
- Using CBOs with new ad sets

The fix:

1. Separate images and videos into different ad sets
2. Use ABOs for high AOV products (forces spend distribution)
3. Wait 14 days before judging based on soft metrics
4. Look at actual ROAS, not just CTR

Your best ad might be your worst problem.

It trains you to stop testing.

Then performance drops and you have no pipeline.

Here's the rule:

Winners get cloned into new angles immediately.

Same proof, new message, new buyer moment.

If one creative drives most of your revenue, you don't have a strategy.

You have a single point of failure.

Why this happens:

You find a winning ad.

It's crushing it.

3x ROAS, scaling well, everyone's happy.

So you stop testing.

Why would you? Things are working.

Then 6 weeks later, the ad fatigues.

Performance drops 40% overnight.

You scramble to find new creative.

But it takes 2-4 weeks to brief, produce, and test new ads.

You have a gap in performance.

Revenue drops.

Panic sets in.

The fix:

Treat every winning ad as a starting point, not a destination.

When you find a winner, immediately ask:

1. What angle is working? (Problem-solution? Transformation? Social proof?)
2. What audience is this speaking to?
3. What buyer moment is this hitting? (Awareness? Consideration? Intent?)

Then clone it:
- Same proof, different message
- Same message, different avatar
- Same avatar, different format
- Same format, different problem

This creates a pipeline of new creative before you need it.

The goal:

Never have one ad driving more than 30-40% of your revenue.

If you do, you're one fatigue event away from a crisis.

Diversify your creative portfolio like you'd diversify an investment portfolio.

Winners get cloned.

Immediately.

Always.

That's how you stay ahead of fatigue.

If you want to reposition a product…

Study what this collagen brand is doing.

1,200+ ads live.
58 videos.
87 images.
UGC.
High production.
Partnership ads.
Dynamic creative.

That’s not chaos.

That’s controlled angle testing.

They’re pushing multiple hooks:

* Weight loss.
* GLP-1 support.
* Skin + hair.
* Bloating.
* Energy.

Here’s the smart bit most brands miss:

Every major angle has its own landing page.

Click a beauty ad?

You land on a beauty page.

Click a GLP-1 ad?

You land on a GLP-1 specific page explaining support, FAQs, context.

That’s how repositioning actually works.

Because here’s the problem:

Collagen does 5,000 things.

If I click a “weight loss” ad…

And land on a generic collagen page talking about joints, hair, gut health and nails…

Continuity is dead.

Conversion drops.

Angle matching is the unlock.

Ad angle → Dedicated page → Reinforced belief.

That’s how you scale multi-angle creative.

Now the opportunity:

They’re running a 28-day challenge.

Great idea.

Challenges are strong because they:

* Add structure.
* Add community.
* Add urgency.
* Increase AOV via bundles.

But the execution matters.

If I land on the page and can’t clearly answer:

“What is this?”
“What do I get?”
“How do I win?”
“Why should I care?”

You’ll leak conversions.

Here’s the fix for any challenge funnel:

1. One clear headline.
   “Join the 28-Day X Challenge to achieve Y.”

2. Bullet the tangible benefits.

* Weekly trackers.
* Expert Q&A.
* Prize breakdown (clearly explained).
* Exact entry rules.

3. Visual roadmap.
   Week 1 → Week 2 → Week 3 → Outcome.

4. Tight FAQ that answers real objections.

Most $5M–$50M brands test more ads.

$50M–$500M brands test:

* More angles.
* More pages.
* More positioning narratives.

Creative volume gets attention.

Angle-to-page continuity gets scale.

If you’re repositioning a hero product this year…

Don’t just change the ad.

Change the destination.

In 2026, creative is the targeting.

But most brands are trying to learn creative the wrong way.

You can find:

* 100s of videos on account structure
* Deep dives on Shopify analytics
* Endless media buying breakdowns

But try finding a clear system for “how to make a good ad”.

You can’t.

Because most creative training is intuition-based.

You’ll hear things like:

“The hook just isn’t strong.”
“The body needs to flow better.”
“This angle feels off.”

If you’re analytical, that’s useless.

You don’t need vibes.

You need a framework.

If creative still feels random in your business, it’s not a talent issue.

It’s a lack of structure.

Here’s a simple 3-step framework we use to pressure test every ad:

1. Attention
2. Communication
3. Direction

That’s it.

Step 1: Does it capture attention?
If they don’t stop, nothing else matters.

Check:

* Is the first 2 seconds visually disruptive?
* Is the first line outcome-driven or pattern-breaking?
* Would you stop for this in your own feed?

Step 2: Does it clearly communicate the product?
Not the vibe. Not the lifestyle.

The product.

Can someone answer in 5 seconds:

* What is this?
* Who is it for?
* Why should I care?

If clarity is weak, conversion will be weak.

Step 3: Does it drive action?
Is there a clear next step?

Not vague.
Not implied.

Clear direction.

* Shop now
* Start your trial
* Get 20% off today

Most underperforming ads fail at one of these three.

Not because the team isn’t creative.

But because they never audited the ad against a framework.

If you’re running $5m–$500m per year, creative can’t be “gut feel”.

It needs to be systemized.

Next time an ad underperforms, don’t ask:

“Is this a bad concept?”

Ask:

Where did it break?
Attention?
Communication?
Direction?

Creative becomes scalable the moment it becomes measurable.

If your ad is “crushing it”…

But you have a big organic audience…

Be careful.

Here’s the influencer ad trap killing scale for $5M–$500M brands 👇

An ad gets:

* 5,000 likes
* Dozens of comments
* Strong blended ROAS

Looks like a winner.

But it’s not the ad.

It’s the audience.

What’s actually happening:

1. You target 100,000 people
2. 50% already know you
3. They convert at 2–3%
4. Cold traffic converts at 0.5% (or worse)

Example:

* $10k to warm → 10x ROAS
* $20k to cold → 1.5x ROAS

Blended result?

$30k spend
$130k revenue
4.3x ROAS

Everyone celebrates.

But you’re scaling the 1.5x.

Not the 4.3x.

The 4.3x is propped up by owned attention.

This is why influencer-led and organic-first brands struggle when moving to paid:

* The hook is the founder
* The creative lacks diversity
* The algorithm isolates into familiar audiences
* Cold never truly performs

Blended numbers lie.

Incrementality doesn’t.

What to do instead:

1. Segment warm vs cold properly
2. Make creative built for cold audiences
3. Expect cold ROAS to be materially lower
4. Run ongoing lift tests
5. Make scale decisions on incremental return, not blended

If you don’t know your incremental ROAS…

You don’t know your real ceiling.

And if you’re pushing spend based on blended MER…

You’re flying blind.

The influencer led brands that scale are the ones who understand where performance actually comes from.

taplio

The Battle

Josh Tay

Nathan Perdriau

Followers

Posts

Engagements

Likes

Performance ScoreJosh Tay's LinkedIn Performance

Performance ScoreNathan Perdriau's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Josh Tay

Nathan Perdriau

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Josh Tay's LinkedIn Performance

Performance ScoreNathan Perdriau's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreJosh Tay's LinkedIn Performance