top of page

Variance: Don't Let It Slap You In The Face

Updated: Jul 20, 2022

Variance is inevitable.

One of the most difficult concepts to grasp, and to plan for, is that of variance. Variance is when life becomes inconsistent and surprises pop up that you did not expect. It can be difficult to deal with because it's completely unexpected (usually) and will throw serious wrenches into your plans.

In marketing, variance can wreak havoc on campaigns, ultimately leading to lower performance and unhappy stakeholders. This can occur when variables completely outside of your control affect your plans.

Facebook may shut down an ad account, a virus may cripple the economy, data laws could drastically change in different regions of the world. It's impossible to be completely prepared for variance in marketing, but there are steps you can take to reduce variance in marketing and help your organization perform better when variance occurs.


Evaluating data for performance is a major component of mitigating variance. If you don't understand why your data has the results it has, then you'll never figure out the right changes to make to optimize performance. That allows variance to sneak under your radar, and potentially damage other areas of your organization down the road. Through correct evaluation, patterns will become evident that will give you data to help reduce variance over time. Here are some ways to help your evaluation process:

  • Patience: Your data needs time to mature and your accounts need time to optimize. We have had clients that become concerned when a campaign does worse than projected and it's only been a week. What do we tell these types of clients? Wait. All campaigns, ALL CAMPAIGNS need a minimum of 28 days to optimize and that is still not very much time. Usually, a good timeline to begin a serious evaluation of a campaign is after 90 days. What does a 90-day timeline provide that shorter time periods don't?

  • A Good Sample Size: Everyone remembers a stats class at some point in their lives in which sample size was stressed. It turns out that sample size IS important, and it looks like you did learn something in math class that you should be using in your real life. Large sample sizes reduce the variance in data and give a better picture of the actual situation. If you would like to get super technical, check out this article from Portland State University.

  • A Good Budget: A good budget actually goes hand in hand with patience. Both time and budget work in balance with each other. If you want to get results quickly (less than a month), then your budget needs to be larger. If you have more time, then your budget can be smaller, and you will probably still be able to accomplish your goals. This is because you still want to capture as much data as possible so that your campaign can optimize, and the ad platform's AI can assist in your campaigns better. The more budget, the more likely you are to reach more people (depending on your ad serving), and the longer the amount of time, the more likely you are to reach more people as well.

  • Time is More Important: Time is the more important of the two (budget and time) simply because it allows you to collect data on TIME DRIVEN variance such as seasonality, cycles in purchase behavior in a month, and market variance.

  • A Good-Sized Target: Most ad platforms allow the advertiser to select very specific demographics and targeting options. Many ad platforms also allow you to add exclusions to your targeting. Although these options allow you to get very specific with your targeting, it's not always beneficial to have a tiny target because your campaign will have a difficult time optimizing. Google and Facebook tell you the estimated market size of your campaign as you put in the different targeting options. Facebook even tells you when it thinks your targeting audience is too small or too large.

PRO-TIP: Facebook has a checkbox called "Detailed Targeting Expansion" that allows Facebook to increase your target audience to an optimal number of people based on correlations between your selected targeting parameters and other targeting options.

Ways to Mitigate Variance

  • Consistency: Consistency is one of the best ways to mitigate variance and bring results in every aspect of life, marketing especially. Making a plan, and sticking to it consistently, will yield results. Consistency takes advantage of time as a factor in reducing variance. Consistency is keeping to a scheduled plan over time. Do not go in "spurts", that is rarely good for your brand as you may start to see results and but once you stop, those results begin to go away. It's just like working out or investing.

  • Reducing the Number of Variables: This tactic is much more difficult to obtain as variables pop up everywhere. Some ways to reduce variables include simplifying your website, having consistent messaging across all platforms, and reducing/ simplifying your offerings.

  • Having a Backup (Diversification): This may seem contradictory to the last bullet point, and that's because it is. We have learned first-hand how important backups are in marketing. One of our client's Facebook Ad Accounts got disabled because of a system error on Facebook's part. The ad account took four days to be manually reviewed and approved. That resulted in an estimated $1,400 loss for those four days. Always make sure that you have multiple ad platforms to operate from (all setup and ready to go.) We use Google Search and Display, Facebook, and now Snapchat. If an ad account goes offline for whatever reason, we always have at least 3 other options to run ads from.

Finance Example:

In finance, there are three big rules for investing:

  1. Dollar-cost averaging: Dollar-cost averaging is an investment strategy in which an investor divides up the total amount to be invested across periodic purchases of a target asset in an effort to reduce the impact of volatility on the overall purchase. The purchases occur regardless of the asset's price and at regular intervals; in effect, this strategy removes much of the variance at specific times. This is similar to our suggestion for being consistent and being patient. It's important to be consistent over time.

  2. Diversification: Diversification is a technique that reduces risk by allocating investments among various financial instruments, industries, and other categories. It aims to maximize returns by investing in different areas that would each react differently to the same event. This is similar to our suggestion on, you guessed it, diversification. If one stock does poorly, another stock may do well under the same circumstances.

  3. Time value of money: The time value of money is the concept that money you have now is worth more than the identical sum in the future due to its potential earning capacity. This core principle of finance holds that provided money can earn interest, any amount of money is worth more the sooner it is received. This relates to our suggestion of patience. Although not as directly as the others, this concept shows that by putting resources to work now, over time, those resources may become more valuable.

How Variance Effects Forecasting

No matter how much time and effort are put into forecasting marketing campaigns, there is always going to be unexpected variance that ultimately makes your forecasts inaccurate. That is why it is important to build in a buffer to all of the forecasts that you make. At Zova Marketing, we usually forecast KPI low to high and use a buffer of about 10%. We're always aiming for the "sweet spot." You can read more about the "sweet spot" and diminishing returns in our article, "Are You Setting the Right Goals?"

EV Mindset

Having an EV mindset when dealing with variance is key. What is EV you ask? Expected value (EV) is the predicted outcome of something. This is one of the main concepts we use to forecast performance for our clients. For example, we can use expected value to find the average revenue a conversion will bring in for a client. We'll have more on that in a future blog.

An EV mindset is key to making good decisions, specifically when you lack complete information (which is 99.9% of the time in life). Here are a few examples:

Texas Hold’em

Poker (Texas hold’em) is inherently a game of incomplete information. When you play the game, you only have knowledge of your two hole cards at the start of a hand. You don’t know what your opponent’s cards are, and you don’t know what cards will come on the flop, turn, and river. Therefore, in order to be a winning player in the game, you must make decisions that have a positive expected value (EV); decisions that will win you money in the long run.

Here’s a very specific example within poker: Say you’re dealt two aces, the best possible hand in poker. The guy to your right goes all-in before any cards are shown on the board (pre-flop). Of course you call because you have the best possible hand, and the other guy ends up having two kings. You’re in great shape to win the hand as an 80% favorite. However, a king comes on the flop giving your opponent three of a kind and you lose. Does that mean you made the wrong decision? Of course not, you had no idea what cards your opponent had (doesn’t really matter when you have the best hand possible anyways) and what cards we’re going to be dealt on the board. Given the information you had, you made the right decision, and a very +EV decision as well. In fact, if you were in that scenario 1,000 times and you made the same decision every time, you would win a massive amount of money because statistically you would win 800 of those hands.


Chess is very different when it comes to information at your disposal. You can see all the pieces on the board at any given time, as well as what possible moves each piece can make. You won’t be able to know exactly what your opponent will do each move, but there is always a move that will get you closer to a win. That’s why a new AI program called AlphaZero has been crushing all the top grandmasters in chess (you can read more about that here). This AI program had proved that the only piece of incomplete information in chess (knowing what your opponent’s strategy is) isn't necessary in winning games, thereby proving chess is a game of complete information.

So why does this matter? Well unlike poker, you’re not making decisions in chess based on EV. You’re making decisions based on whether the move is winning or not, because you should be able to tell (with some practice of course) if a move will get you closer to winning or losing in each specific position you're in. In chess, it pays to have a results oriented mindset for getting better. If you’re playing someone and you miss a checkmate they have against you, that’s completely on you. You had all the information showing it would happen at your disposal and you didn’t properly leverage that into making the best move. Whereas we were evaluating decisions based on expected value in chess, we're evaluating decisions based on each moves result in chess, and it's all due to the difference in information at our disposal.

Developing the Mindset

When it comes to having an EV mindset, the key takeaway is that whenever you have incomplete information in a situation (which is the vast majority of the time), judge decisions based on their expected value at the time you made the decision rather that the result it produced. The more variance / less information you have, the more you should rely on EV for evaluating your decisions. This is key to improving your decision making process and safeguarding yourself from regret.

The best you can do is make the most +EV decisions you can because eventually, in the long run, you'll win big.

Watch our show, Marketing Tip Tea Time at 2:22 and our episode on variance at the link below! (Josh and I also have super awesome mustaches.)

How do you deal with variance and the unexpected in your business? Let us know in the comments below!

21 views0 comments


bottom of page