👋 Hey, it’s Sundar! Welcome to experiMENTAL: a weekly newsletter on B2C marketing & data science frameworks, how-to guides, and stories from 15 years including early Uber.

Exciting alert! I’m now on Bluesky. click on the section below to follow me.

Why Marketing targets are overrated.

Welcome to 2025. 60% of you are still waiting on targets. 30% of you have targets that make zero sense but you had to put together because you know… 2025. 9% of you have targets that you know will change next week. That leaves 1% of teams that have reasonable Marketing targets built on time that they can start to execute on.

I don’t have a theoretical issue against targets. I have a practical one formed by 15+ years of target setting at startups, large corporations, and the government. I have yet to see targets built correctly, measured consistently, or achieved efficiently. There is always an issue.

Below, I’ll explore the various challenges I see with the entire process of Marketing targets but I’ll also share a better way to approach them.

Targets vs Forecasts

What is the difference between a target and a forecast?

To me, a forecast is what we believe will happen given all of the information we know today. A target is what we want to happen given what we set to do at the beginning of the year. It’s a goal.

What we believe will happen and what we want to happen are never the same.

Yet, companies mix the two. The worst is when you take a forecast and add a multiplier to bump it up and call it a target.

At the beginning of the year, forecasts and targets should be the same. If they’re not, then you might as well throw out targets and not use them. But let’s move forward a quarter. You are 10% behind Q1 targets. This means there was some reality that didn’t comply with your assumptions.

You carry forward this underperformance forecasting the rest of the year will be down, but targets stay the same. What number should you look at? It’s nice to say “We’ll make up our targets next quarter”, but how often does that really happen? If you missed targets first quarter, why do you believe you’ll start overperforming? Was this an issue in target setting or business performance?

You then spend the next 3 quarters sharing why forecasts were behind targets. So fun.

A waste of resources

General process to set targets:

You get an email from your manager in early November “Hey Bob, it’s time to start building 2025 targets!”

Every person scrambles to:

1. Put together a budget and Marketing plan

2. Present them in a new document the new CMO put together

3. Redo it because of arbitrary feedback from a committee of peers

4. Redo it again because finance came back and said “We need to find $XM in the budget”

5. Repeat until everyone pretends to be happy

This process takes months. Imagine how much more Marketing you could get done if you actually spent those weeks doing Marketing. Crazy.

Tops down vs bottoms up

Every good model is a blend of tops down strategy and bottoms up data driven assumptions except Marketing targets which feel mainly top down.

“We need more. We need to show more. We need to do more. Oh and we cut resources”. Regardless of what the data shows, there’s always pushback.

This happens so much that the original forecasts and the end targets are worlds apart. It defeats the whole exercise of the original bottoms up plan.

You know what they say when you assume…

Targets are one giant delicious stew of assumptions. Humans are notoriously bad at assuming what’s going to happen in the next week let alone a year from now. It’s not always that we over estimate. We under estimate too.

CPMs are a great example of assumptions that make or break Marketing targets:

Looking at the chart above, TikTok CPMs went up 15%, Meta 2%, and Twitter down 75%. What’s the probability that you accurately modeled this? You might look at Twitter and say CPMs went down hoorah! Well why did they go down? Because X turned into a pile of right wing crap and advertisers and people left it. I highly doubt you predicted that.

CPMs will go up every year but how much? If CPMs go up by 5% vs what you expected then your targets will be off by 5% no matter what you do. You have to make up for that because of a forecasting error.

That’s one assumption for one channel.

Ownership is an illusion

Ownership of targets is complex because businesses are intertwined and complex.

Example: Paid Acquisition Targets

You’re a Paid Acquisition manager and are asked to build targets for 2025. You start with a top down view and build out a projected # of signups.

In a typical Paid Acquisition funnel, there is 1 metric you actually own and fully influence and yet you are responsible for the entire # of target signups. Ownership incentives are misaligned and not representative of reality.

With any Marketing metric that is a business goal, you have to intersect with Product + Eng.

“Okay, Sundar but then why don’t we make goals we own?”

Great question, Sangita. Because those metrics don’t show ROI and then Marketing gets accused of not driving value for the business. If you’re a lifecycle Marketing team and you set your goal as 100 emails sent you would be fired the next day.

It’s a lose lose.

Strategy is a moving finish line

Simply put, strategy is an ever evolving organism within an organization. There are new product, new audiences, new bets, and competitive changes that make it difficult to know what to prioritize. That’s the reality for any company especially startups / scale ups. This makes setting targets virtually impossible.

Looking a step further by business maturity, there’s never a place where it makes sense:

Early Early stage → too much uncertainty

Early stage → not enough data

Middle stage → not stable / consistent enough data

Late stage → too many new variables (geos, products, use cases)

Late late stage → add a x% growth rate on last year. No need for a complex process.

Psychology is a bit*h

The psychology behind targets is full of pitfalls and traps and it’s the biggest reason I dislike targets.

Studies show that overconfidence bias affects 70% of people when setting goals. We always underestimate challenges and overestimate their abilities.

Then, when you’re behind targets you find a reason to explain why and people only remember that you made an excuse and not the reason. This is because of anchoring bias.

It’s inevitable and you spend the rest of your time explaining yourself and overcoming this perception of “You missed targets”. Your credibility is anchored to the targets you set which as we know are not always rooted in reality.

Setting targets is still the age old way to motivate your team. Find a better carrot or a better team.

How Uber did it

I usually have great things to say about Uber, but not today. The bigger the company the bigger the sh*tshow. .

Performance Marketing → While I was on Performance Marketing team, we didn’t have business targets. Instead we had a ROAS target. These were set in conjunction with P&L owners on the operations team. There were 2 key challenges with this approach:

1. ROAS as a metric →I won’t go deep onto why ROAS isn’t a great metric but it’s not. Primarily because the “Return” in ROAS is hard to quantify. In addition, it uses revenue and not a more relevant like profit which accounts for quality of revenue

2. Ownership of ROAS → Marketing owns the denominator. Marketing + Product + Ops own the numerator. You know how I feel about ownership.

We also had forecasts and when they deviated, we dug into why. Not hitting forecasts is what prompted us to go deeper on figuring out if our Meta spend was incremental (turns out it wasn’t. read more here).

Brand Marketing → This was much more intense of a process and started in November. Teams were asked to write their goals, budget asks, and strategy. They had to be one page and then country managers reviewed it with region managers who reviewed it with the CMO and the loop would continue until people were satisfied. That didn’t happen until end of January. As you can imagine as a Brand manager, you then need to spend time planning and activating so by the time you have a campaign out it’s closer to March at the earliest.

Business goals varied from more bottom funnel to upper funnel depending on the market and strategy so it was hard to measure. Honestly, I’d say it was a mess but that’s how I’ve always seen it. Now that I’ve had the time to step back, I know there’s a better approach.

How to build better Marketing targets

In this example, we’re going to step by step build a Marketing forecast as a CMO responsible for owning signups. Here we go.

1. Eliminate the term target

   I advocate for a complete elimination of the term target. It’s a sales term and should only be used in sales.

   For all other functions, it should be called forecast. A forecast suggests a quantitative and scientific approach. If it was based on qualitative hunches, it would be called a prophecy.

   

   The only prophecy I care about. IYKYK.

2. Align on a singular primary business metric of success

   You need to select a singular primary business metric of success. As CMO, you are told signups is your primary business metric of success so you put out the statement: “We forecast we can hit 100 signups. We intent to stay under our allotted budget of $X”.

   If you get pushback that you MUST stay under $X then $X is truly the primary business metric.
   
   Your new statement is “We forecast we can stay under $X. We intend to deliver 100 signups.” Notice the difference. It’s very powerful.

   
   By selecting the primary metric, you also select the “tops down” strategic focus that is required to build a strong forecast.
   

3. Build a well documented BOTTOMS UP forecast

   Bottoms up forecasts are the only way to make it work. They take 2 paths:
   Budget first → Start with the budget and back into how many signups you can deliver

   Signups first → Start with the signups you’re asked to deliver and then build backwards to the budget you need.
   
   Either way works and should lead to the same answer. In addition, document the sh*t out of your assumptions. Here are some examples:
   
   “Industry research suggests CPMs will go up by 8%. Absence of better assumptions, we will incorporate this increase in CPMs”
   
   “We ran a test last year and believe that we are hitting diminishing returns in our cost curve. We estimate that increasing budget by 50% will result in signups increasing by 35%. See test here”

   
   “We have not incorporated any additional increase in CRO. We have spoken with product and they do not have a good estimate of impact given resource constraints.”
   
   Don’t underestimate the number of assumptions you’re going to make: activation rates, retention rates, LTV, basket size, etc. These are all Marketing metrics too.
   
   Now the forecast from the CMO looks like “We forecast we can hit 100 signups. We intend to stay under $400 assuming a $5 CPM (baked in increase of 8%) and 3% website conversion rate with no improvement in CRO this year .”
   

4. Set expectations on ownership

   Take a good hard look at your assumptions you made in step 3. For each assumption, assign a category of Fully Own, Co Own, External. Don’t be tempted to think you own something that’s external like CPM. You own the relative CPM levels (narrow vs broad targeting, interest based, etc.) but you can’t own the CPM levels themselves. Those are fully outside of your control.

   
   For each co owned assumption, go to the corresponding team and agree that this metric is co owned.

5. Widely share the forecast

   The delivery of the forecast is a crucial step that is often overlooked, Because you’re deep into the forecast doesn’t mean everyone else is. Often a major criticism of forecasts is that they are black boxes full of arbitrary assumptions and opaque decisions. You can change this perspective by setting up a meeting where you go deep on the forecast. People will have their own perspectives and challenge assumptions. That’s a good thing. Better now than later.
   
   You have ONE opportunity to get this right. Deliver to all levels of your team and across silos. Get this out there.

6. Set up the correct rituals
   

   Weekly check ins → Surprises are what demotivate teams and piss off management. To avoid surprises, increase the frequency of check ins. Performance vs forecast needs to be discussed weekly. If the gap between forecast and performance is wide then you need to understand if forecasts were wrong or performance was wildly off. For example, RIFs cut 25% of your team. You HAVE TO bake that into your actual performance. The best way to do this is by reforecasting.

   
   Reforecasts → Depending on the maturity of your business, you should reforecast anywhere from 1x a month to 1x every 6 months. No business can forecast for a year.
   
   A reforecast takes in the actuals as the new assumptions along with a post mortem on why actuals vs forecast were wrong. This is where the ownerships step really comes into play. If as CMO you did everything right but product didn’t ship on something it said it would then are you really to be blamed? It’s not about the blame game. It’s about accountability.

That’s it. In the end, it’ll be similar amounts of work as most teams are already actioning some form of these 6 steps, but making each step more efficient makes the whole process more efficient. It leads to more effective outcomes and less frustration.

Missed my last newsletter?

Don’t be sad. Here it is: Sundar Unwrapped: 2024 in review