March 30, 2023

One of the challenges faced by Catalog Brands is dealing with a Gift Buyer. A Category Development project shows that these customers are unlikely to purchase the rest of the year.

Pretend that Gifts are 10% of a July catalog ... while all other merchandise is 90% of the catalog.

Pretend that we have a customer with a Gift QuickScore of \$70.00 and a General Merchandise QuickScore of \$30.00. Pretend that Gross Margin Percentages are 50% for all merchandise.

• Catalog QuickScore for Gifts = \$70.00 * 10% of Catalog * 50% Margin = \$3.50.
• Catalog QuickScore for All Other Merch = \$30.00 * 90% of Catalog * 50% Margin = \$13.50.
• Total Catalog QuickScore = \$3.50 + \$13.50 = \$17.00.
Now, pretend we have a customer with a Gift QuickScore of \$8.00 and a General Merchandise Quick Score of \$92.00. Same Gross Margin Percentage of 50%.

• Catalog QuickScore for Gifts = \$8.00 * 10% of Catalog * 50% Margin = \$0.40.
• Catalog QuickScore for All Other Merch = \$92.00 * 90% of Catalog * 50% Margin = \$41.40.
• Total Catalog QuickScore = \$0.40 + \$41.40 = \$41.80.
Mind you ... both customers are exactly equal in terms of quality ... but "what" you offer paired with "what" the customer likes to purchase results in two customers with two very different levels of value.

Seems like this is something you'd be interested in knowing, right?

March 29, 2023

Catalog Brands: Applying QuickScores to Category Development

Let's assume you have five (5) Merchandise Categories ... we'll assume something simple.
• Mens
• Womens
• Kids
• Accessories
Let's assume that the gross margin percentages for each category are predicted as follows:
• 60% for Mens
• 62% for Womens
• 55% for Kids
• 50% for Accessories
Ok. For many of you, I create QuickScores for each merchandise category. The QuickScore tells us how much you'll spend in each category. Now let's assume you are a Catalog Brand, and you have the following percentages of merchandise being offered by catalog for your July catalog.
• 20% Mens
• 30% Womens
• 10% Kids
• 20% Accessories
Yeah, lots of percentages.

Pretend we have the following QuickScores for one customer ... one that loves Accessories:
• \$20.00 for Mens.
• \$30.00 for Womens.
• \$5.00 for Kids.
• \$60.00 for Accessories.
What is an appropriate QuickScore for this Catalog?
• Mens = \$20.00 * 20% Mens * 60% Gross Margin = \$2.40 of Margin.
• Womens = \$30.00 * 30% Womens * 62% Gross Margin = \$5.58 of Margin.
• Kids = \$5.00 * 10% Kids * 55% Gross Margin = \$0.28 of Margin.
• Gifts = \$5.00 * 20% Gifts * 35% Gross Margin = \$0.35 of Margin.
• Accessories = \$60.00 * 20% Accessories * 50% Gross Margin = \$6.00.
• Gross Margin Score = \$2.40 + \$5.58 + \$0.28 + \$0.35 + \$6.00 = \$14.61.
Now, pretend we have the following QuickScores for another customer ... one that loves Womens:
• \$10.00 for Mens.
• \$90.00 for Womens.
• \$10.00 for Kids.
• \$5.00 for Accessories.
What is an appropriate QuickScore for this Catalog?
• Mens = \$10.00 * 20% Mens * 60% Gross Margin = \$1.20 of Margin.
• Womens = \$90.00 * 30% Womens * 62% Gross Margin = \$16.74 of Margin.
• Kids = \$10.00 * 10% Kids * 55% Gross Margin = \$0.55 of Margin.
• Gifts = \$5.00 * 20% Gifts * 35% Gross Margin = \$0.35 of Margin.
• Accessories = \$5.00 * 20% Accessories * 50% Gross Margin = \$0.50.
• Gross Margin Score = \$1.20 + \$16.74 + \$0.55 + \$0.35 + \$0.50 = \$19.34.
These are two customers that have exactly equal future value ... but one customer skews to Women, and the upcoming catalog is also skewed to Women. Therefore, the latter customer is much more valuable.

When we develop categories, we make conscious choices. Sometimes we use models to determine who to send catalogs to. Sometimes those models have an "overall" score that does not take categories into account. If we align categories with gross margins per category and space allocated within the catalog, we obtain a different outcome.

March 28, 2023

Unprofitability of the Undifferentiated Middle

Here's what "the undifferentiated middle" looks like when you evaluate metrics down to profitability at a customer level.

There are problems when you increase price points to cover cost of goods increases.

Here, we see that rebuy rates decreased from 35.0% to 32.5%. That's a consequence to increasing prices to cover cost of goods increases.

We see that units per repurchaser decreased from 5.00 per year to 4.57 per year. When the customer purchases, the customer buys fewer items to compensate for higher prices.

We see that the price per item purchased increased from \$30.00 to \$34.50 ... covering the cost of goods sold per item from \$12.00 to \$16.50. In this case, the brand simply passed the cost increase along to the customer.

However, the mix of fewer units times a higher price yields an increase in spend ... from \$150 per year to \$157.50. So yes, this brand got those who purchased to spend more. Unfortunately, fewer customers purchased ... so the math is about to fail.

Under the 2021 business model, we generate \$52.50 of demand per customer ... under the 2023 business model we generate \$51.19 of demand per customer (rebuy rate * spend per repurchaser gets us to \$51.19). The higher cost of goods sold sinks the p&l, leaving us with \$9.93 of profit today vs. \$14.13 two years ago.

This is where the undifferentiated middle leaves us when we have a cost of goods increase.

No loyalty program fixes this.

A robust customer acquisition program fixes this challenge ... but you need a lot more new customers than you initially think.

This is why we turn to categories ... we have categories that yield more profit than average and have minimal cost of goods increases. Those become categories we have to feature, especially among new customers who may become attached to those categories going forward.

March 27, 2023

World Going in Three Different Directions

People grumble about dollar stores ... but those stores are the logical extension of the Walmart/Target models. If Walmart found a way to structure the Price/COGS relationship to their advantage, it makes logical sense that somebody else would find a way to use the Price/COGS relationship against both grocery stores and Walmart.

Retail and e-commerce move in three different directions.

1. Lowest Possible Price (dollar stores, Walmart, Amazon).
2. Highest Possible Gross Margin Differential (fashion / experiences).
3. The Undifferentiated Middle (most of us).

You are going to tell me you cannot run a business at the lowest possible price.

You are going to tell me you cannot run a business that is fashion-centric with great experiences.

This means you are going to have challenges. It doesn't mean you cannot be successful, but you are constantly pressured by (1) and (2) above, on opposite ends of the spectrum. It's like a rubber band that you pull on each end until it breaks in the middle.

Tomorrow we'll look at a profit-and-loss statement at a customer level, showing what "the undifferentiated middle" looks like.

March 26, 2023

Dollar Stores

You aren't going to read about or hear a lot about dollar stores in the retail gurusphere.

The price point / cost of goods relationship is generally overlooked by the retail gurusphere. It isn't overlooked in a Category Development project. There are clear relationships identified in each project, and profitability is ultimately defined by the relationship.

More on the relationship tomorrow.

March 23, 2023

The DTC Index - And Giving Away Free Information

Pay attention to what the author says about Marginal aMER. In my simulation / forecasting work, the relationship described comes through when measuring profit per new customer (though the relationship isn't identical to what is illustrated here).

P.S.:  A reader recently responded that he wanted "more free stuff". If you are that reader, click on the link and subscribe to their free stuff.

P.P.S.:  This is interesting. There is a clear relationship between giving away free insights and me staying in business. The relationship looks something like this ... the X-Axis represents the percentage of my blog posts where I give away my methodology for free ... the Y-Axis represents the percentage of annual income I generate as a result.

In other words, if I give away my methodology in 20% to 40% of my posts, I maximize my annual income. Give away everything in every post? I lose about half of my annual income. If I give you more free information, I eventually make less income and I'm out of business. Such is the relationship between a business and customers. Always remember that not all customers are customers ... some are consumers ... and there is a big difference. You make your profit off of customers, you generate improved vanity metrics from consumers.

March 22, 2023

We've talked about your Category Development situation as being similar to a solar system ... you have a Star, you have Planets that rotate around the Star, you have Moons that rotate around Planets, you have Asteroid Belts (broken categories), and you have Comets that infrequently interact with your solar system (i.e. Gifts at Christmas).

And when you fail in one category, you impact other categories.

If your Star category struggles, the struggles spill over into all other categories.

If a Planet category struggles, the Moons will struggle.

If a Comet category struggles? Meh.

But make no mistake. Your business is interconnected, and you are doing things that greatly help or hurt other categories.

March 21, 2023

When you are developing categories, you likely look at your cost of goods and determine a price that (theoretically) delivers enough profit to fuel your business into the future.

On average (your mileage will vary), when prices increase, customers decrease. The goal (on the surface) is to increase sales as prices increase ... you trade off a few customers but the math all works out.

Here's what you are hoping to accomplish.

• Last Year = 1,000 customers buying 1.4 units per customer at \$50 each = \$70,000.
• Next Year =    800 customers buying 1.3 units per customer at \$70 each = \$72,800.
• Sales Increase = 4%.

That's what you are hoping to accomplish.

But there is a downside.

With 200 fewer customers, you have 200 fewer customers who will buy other merchandise from other categories. This creates a feedback loop - with fewer customers available to buy from other categories, those categories struggle, you end up with too much merchandise in those categories, you have to liquidate that stuff at lower prices (which increases customer counts but at a purchasing disincentive which hurts you long-term), and profit is hurt.

In other words - be careful! As prices increase, customers decrease, leaving you with fewer customers to buy from other categories.

March 20, 2023

Three Ways To Grow

1. Get more sales from loyal buyers.
2. Acquire new customers.
3. Manage your merchandise assortment properly.

Rank-ordered by what my project work suggests is most important?
1. Manage your merchandise assortment properly.
2. Acquire new customers.
3. Get more sales from loyal buyers.

March 19, 2023

How Many New Items?

Ok, here's our table.

There is a relationship hidden in the table.

• 594 items = \$8.0 million in new item sales.
• 235 items = \$3.9 million in new item sales.
• 579 items = \$9.5 million in new item sales.
• 603 items = \$8.7 million in new item sales.

Graphically, the relationship looks like this:

Here we have our friendly law of diminishing returns ... it appears practically everywhere once you know how to look for it.

The first 100 new items generate about \$2.0 million in sales.

The next 100 new items generate about \$1.5 million in sales.

The next 100 new items generate about \$1.4 million in sales.

The relationship progresses from there. The first three hundred new items get you just shy of \$5.0 million ... while the next three hundred new items add about \$3.8 million.

If you needed to get ten million in sales from new items next year, you'd need 700 new items next year.

So yeah, this stuff is really useful!

March 16, 2023

What Harms You Today Harms You Tomorrow

We learned yesterday that the Class Of 3/10/2021 was too small, and didn't generate enough sales, costing this business five million dollars in the year ending 3/10/2021.

Did you notice that the business never recovers from this problem?

New items recover quickly. Somebody noticed that the merchants nuked the business, and as a consequence new items perform at \$9.5 million the following year ... recovering by about \$5.5 million dollars. However, the total business does not recover ... sales increase by a paltry \$0.4 million.

How can that be?

Two things are happening here.

First ... new items today become existing items tomorrow. Read across the year ending 3/10/2021 row. \$3.9 million in year one becomes just \$2.6 million in year two. In the class of 3/10/2020, \$8.0 million in year one becomes \$5.7 million in year two. In other words, this brand loses \$5.7 million - \$2.6 million = \$3.1 million in sales in the second year because the merchants failed the year prior.

Each merchandise class has a "life" ... and if you don't have enough new items one year you won't have enough existing items the following year.

That's the first issue.

The second issue? This is a completely different issue. Look at the top row. These are older items, items that have been around for more than four years. Look at how sales drop off year-over-year. These items gave up \$6.6 million in the year ending 3/10/2020, they gave up \$6.8 million in the year ending 3/10/2021, they gave up \$5.6 million in the year ending 3/10/2022, and they gave up \$3.0 million in the year ending 3/10/2023.

In other words, items at this company have a moderate life cycle (in terms of length), and each year you need to come up with enough new items to cover the losses you experience from long-term existing items. If you don't generate enough new items, you hurt your new item performance AND you fail to compensate for items that are aging.

You can use this analysis technique to measure how many new items you "might" need next year to protect the future of your business. When you know the decay rate of existing items you can back into how many new items you likely need to keep your business moving forward.

March 15, 2023

Problem Spotted!

Back to our table:

Look at the row titled "New Through Year Ending 3/10/2021".

That "class" had just 235 new items, compared to 594/579 in the years before/after it.

This merchandise class was small, and it did not generate sufficient sales. In the first year, this "class" generated \$3.9 million in sales, a lot less than the years before/after it (\$8.0 million, \$9.5 million).

Look at total sales for the year ending 3/10/2021 ... there is a five million dollar top-line hit compared to the year prior.

This business was nuked by a merchandising team that failed to give proper importance to finding new items that work. Some might say "well, we had supply chain issues" and sure, that's a problem. But it doesn't solve the problem of the business being harmed.

Notice that this business doesn't recover, with sales down for each of the two subsequent years. More on this topic tomorrow.

March 14, 2023

Can You Spot The Problem?

Though this seldom happen in actual project work with actual clients, there are times when I receive emails or tweets that share this theme:

• "You seem to oversimplify everything. Business really doesn't come down to new customers and new merchandise."

Well, yeah, it does come down to that.

Marketing experts generally accept that customers have long-term value. If you acquire a customer today, the customer will buy stuff tomorrow ... at ever-decreasing rates of return of course, but future purchases will happen. Future purchases pay for today's acquisition costs.

Very few marketers understand that their marketing performance is directly influenced by the outstanding or unacceptable merchandise performance they have no control over. Maybe your online conversion rates average around 3% and over the past six months they've slumped to 2.7%. Who gets blamed? The marketer! It must be the traffic. The marketer made a mistake.

Often, it's the merchant who messed up - but nobody can see the truth because nobody reports on new merchandise performance. Just as often, the issue is easy to identify and not hard to fix.

Here's our table from yesterday.

Each year there is a new merchandise class. Those items sell for several years before being discontinued and/or sales decline.

Look at each merchandise class.

Can you spot the problem?

March 13, 2023

The "Class Of" Report

Over the past two weeks I mentioned that I was going to bring back a table from Merchandise Forensics work I performed a decade ago. This table became really important (again) in 2021 when supply chain issues impacted our businesses.

The table is called a "Class Of" table. In it, we review new items introduced by year ... each year is a "Class", and we follow the "Class" into the future to see how sales change/evolve.

Here is an example ... review it today, we'll discuss it tomorrow.

March 12, 2023

Silicon Valley Bank

I don't have any insider information, sorry. Here's an article to help you learn a bit more.

Business Professionals struggle with feedback loops. Feedback loops are cousins of interactions. Here is an example of an interaction. This is a 2x2 factorial test on conversion rates.

Here we see that the interaction of B/C produces the best result. A test of this nature identifies the interaction we should pay attention to.

Now, what would happen "A" was your control and you tested A vs. B? The results would be identical, and you'd conclude to keep doing "A" and you'd never realize that A with B works best.

In our world of A/B testing, we miss interactions all the time. Interactions lead to unique outcomes, and unique outcomes can lead to feedback loops that can either accelerate results or harm our businesses.

I wrote this more than eight years ago for catalog readers (click here). Here we are in 2023 with constant grumbling about how hard it is to find new customers in large enough quantities. With an inability to measure and understand interactions between Amazon and online brands and catalog brands (all of which could have been disclosed by the co-ops but they'd have seen that as bad for their clients), my catalog clients suffered mightily, and continue to suffer.

In the case of Silicon Valley Bank, you saw what the interaction of risk (which likely wasn't as huge as it was on many/most bank failures) did when multiplied by insider Silicon Valley folks pulling money early multiplied by commoners pulling money late leading to the failure of the bank.

Where possible, develop your skills at identifying interactions. It's a critically important business skill to acquire.

March 09, 2023

Retail Stores

Here's a quote I now use on Twitter.

• "A store is a purpose-built structure designed to extract money from customers and prospects who want to be there."

What's coming for retail is, on one hand, going to be messy and unpleasant. But on the other hand, what retail professionals currently under the age of 45 are going to do to reinvent retail ... that's going to be amazing! Smaller stores aren't the answer, they're a step in getting to the answer. The movie theater article above? That's a parallel to turning a store into a distribution center ... it's not the answer, it's a step in getting to the actual future.

The actual future? I'm excited about it!

March 08, 2023

Odd Quirks in Category Targeting Models

Here's one of the odd quirks you find when creating Category Targeting Models. This brand has twelve (12) main categories. When I decile my findings, I see oddities. Look at this customer (decile 10 is best, decile 1 is worst).

• Category 01 = 10.
• Category 02 = 10.
• Category 03 = 10.
• Category 04 = 8.
• Category 05 = 10.
• Category 06 = 10.
• Category 07 = 10.
• Category 08 = 10.
• Category 09 = 10.
• Category 10 = 10.
• Category 11 = 10.
• Category 12 = 10.
You'd want to target this customer with EVERY category, correct? Yeah! What's interesting is that this customer possesses weighted dollars within Categories 1/6/7/8/9/10/11. That's a really good customer. And all of that cross-shopping suggests that the customer would also be a prime candidate for Categories 2/3/4/5/12, even with no prior purchase history in those categories.

This is not an uncommon outcome.

Your best customers should be exposed to EVERYTHING you sell, period.

It's the marginal customers where targeting makes a huge difference. Look at this customer.

• Category 01 = 2.
• Category 02 = 2.
• Category 03 = 3.
• Category 04 = 4.
• Category 05 = 2.
• Category 06 = 4.
• Category 07 = 4.
• Category 08 = 2.
• Category 09 = 2.
• Category 10 = 7.
• Category 11 = 2.
• Category 12 = 2.
There you go ... target that customer with Category 10, right?

Each brand possesses some odd quirks in their Category Targeting Models. Figure out what those quirks are, and when you are targeting your customers take advantage of the odd behaviors you observe.

March 07, 2023

Category Targeting Model Framework

I've explained this previously, and the statisticians in the studio audience are going to belly-ache about what I'm about to explain, but here's what I do ... I'm transparent about my methodologies.

Step 1:  Select twelve-month buyers as of 12 months ago.

Step 2:  Calculate historical spend in 0-12, 13-24, 25-36, and 37-48 month buckets.

Step 3:  Calculate future spend in the next twelve-months.

Step 4:  Regress the four variables in Step 2 against the dependent variable in Step 3.

Step 5:  Evaluate the coefficients for the 0-12 month spend variable, 13-24 month spend variable, 25-36 month spend variable, and the 37-48 month spend variable. Let's pretend that they are as follows ... 0.604 ... 0.369 ... 0.235 ... 0.134.  Divide each value by the value of the 0-12 month coefficient (0.604), giving us 1.000, 0.611, 0.389, 0.222.

We now have our weights ... we weight 0-12 month dollars by 100%, we weight 13-24 month dollars by 61%, we weight 25-36 month dollars by 39%, and we weight 37-48 month dollars by 22%.

Step 6:  For each category, we calculate weighted dollars spent during the past four years, as of one year ago.

Step 7:  For each category, we calculate total dollars spent in the past year.

Step 8:  We run a regression for each category ... the independent variables are weighted dollars spent in each category, the dependent variable is amount spent in the past year for a specific category. This way, we learn which categories contribute to future spend within a specific category.

Step 9:  We repeat Step 2, shifting our date ranges one year forward.

Step 10:  We score each customer for each category for the next year.

Step 11:  If desired, results are placed into deciles for each category.

At this point, you have completed Category Target Modeling ... you now have targeting variables for every category. If you wish to send an email campaign to Mens Outerwear customers, you use the Mens Outerwear Category Targeting Model results.

Again - statisticians are gonna grumble about the methodology. Let them come up with their own ... this methodology is simple and effective and proven to work in the real world.

P.S.: Yes, you can use logistic regression here as well with a 1/0 dependent variable. Works well.

March 06, 2023

Category Targeting Models

If you want to develop your categories, you'll need two things.

1. A style/sku plan to properly manage winning items, new items, and existing items.
2. A marketing plan to target merchandise to customers most likely to purchase from various categories.

Next week we'll talk about "Class Of" reporting, reporting that helps us see what problems we have within various categories.

This week, we'll talk about Category Targeting Models. All of you should have Category Targeting Models, and when you feature anything you should feature it to the customers most likely to purchase it (though exposure is important as well).

Tomorrow, I'll share with you my Category Targeting Model Framework.

March 05, 2023

Loyalty Program Expectations

There is a reason so many retail brands tie loyalty programs to credit programs. The goal isn't necessarily to increase customer spend - the goal is to increase the amount of interest charged to a customer.

In my work, it is generally true (your mileage will vary) that loyalty programs increase customer spend by around 10%, plus/minus.

For a loyalty program to work, you need high value customers. Pretend you have a typical e-commerce customer with a 30% chance of buying again, purchasing 1.6 times at \$100 each if the customer repurchases.

• Normal Conditions = 0.30 * 1.6 * 100 = \$48.00 of expected spend next year.
• Loyalty Program = 10% Bump = \$48.00 * 0.10 = \$4.80.

It doesn't take a rocket scientist to notice that increasing spend by \$4.80 a year per customer in a loyalty program won't do anything, after you subtract all normal business expenses and then subtract marketing expenses associated with a loyalty program.

Take a customer with a 70% annual rebuy rate, 5 purchases per year at \$100.
• Normal Conditions = 0.70 * 5.00 * 100 = \$350.00 of expected spend next year.
• Loyalty Program = 10% Bump = \$350.00 * 0.10 = \$35.00.

You can make an argument that \$35.00 is a credible amount of increased spend, so yeah, have at it, create a loyalty program for this customer.

The problem, of course, is that you don't have many of these customers.

On average, a customer needs to achieve a fifth purchase before the customer has a 60% or better chance of repurchasing next year. Once a customer has a 60% chance of buying next year, the customer generates sufficient net sales to make loyalty programs both noticeable and meaningful (see my first example above for a situation that is not noticeable and is not meaningful).

Here's an exercise you can perform with your own data.  Segment all twelve-month buyers by number of life-to-date purchases as of one year ago today. Select all customers with 5+ life-to-date purchases from that audience. Then calculate total spend in the next year from that audience. Finally, multiply that number by 10%. That's the expected amount of sales increase a credible loyalty program "might" deliver. Again, your mileage will vary.

For a typical e-commerce business, the audience of 0-12 Month 5x+ buyers generated maybe 20% of annual sales, plus/minus. Multiply that amount by 10% and it means a credible loyalty program "could" add 2% to annual net sales levels.

This is why large retail brands tie loyalty to credit ... it allows them to make profit off of the interest.

This is why I advocate for e-commerce brands to simply go find another new customer ... the math works out so much better over time.

March 02, 2023

Statistical Significance of A/B Tests

There is a ton of misinformation out there about A/B testing. Those lacking rigorous statistical training tell you that you need "x" responses for a valid A/B test.

That's not how this stuff works.

More than thirty years ago (Lands' End), we developed an equation to estimate the variance associated with our A/B mailing tests. It turned out that the variance of our estimates was non-constant. In other words, the variance might be "x" when the dollar-per-book was \$3.00 ... it might be "1.5x" at a dollar per book of \$5.00.

We developed an equation ... as long as dollars-per-book was >= \$2.00 variance could be estimated as -188 + 192*x, where "x" was the dollars-per-book in a test group (aside ... there are going to be statistical experts who balk at creating this equation ... one that accounts for non-constant variance ... and will say that everything that follows is garbage ... just want you to know that view is out there, I need to be forthright here).

Let's pretend that our control dollar-per-book was \$3.00, and we expected the test dollar-per-book to be \$3.25. Let's pretend that we wanted 10,000 customers in the test group and 10,000 customers in the control group. Would our results be statistically significant?

The t-test equation looked like this:

• Test Group Dollar-Per-Book = \$3.25.
• Control Group Dollar-Per-Book = \$2.75.
• Test Group Sample Size = 10,000.
• Control Group Sample Size = 10,000.
• Variance of Test Group = -188 + 192*3.25 = 436.
• Variance of Control Group = -188 + 192*3.00 = 388.
• T-Score = (3.25 - 3.00) / SQRT(436/10000 + 388/10000) = (0.25) / (0.29) = 0.86.

The T-Score is nowhere close to 2.00 ... so the results are not statistically significant.

Now, does that mean that the results aren't meaningful? Maybe. What would happen if we had 100,000 customers in each of the test/control group?

• T-Score = (3.25 - 3.00) / SQRT(436/100000 + 388/100000) = (0.25) / (0.09) = 2.78.

Now the results are statistically significant.

What was the difference?

Well, the original sample size was too small.

We used the equation above to determine the appropriate sample size for all tests based on the amount of variance associated with our expectations for test group performance and holdout group performance.

When I worked at Nordstrom, we used a comparable equation - one specific to Nordstrom. We learned that we needed 100,000 or 200,000 customers to measure what we wanted to measure ... not 10,000 or 20,000.

Comparable issues impact website conversion. You don't want to measure conversions ... you want to measure sales per visitor. Your test group might spend more/less than the control group once the decision to purchase is made ... so you have to measure sales per visitor.

March 01, 2023

There Goes Google ... A Channel All By Itself

In a recent project, here's what the data told us about Google.
• 7.8% of sales were derived from one specific Category.
• 20.4% of all sales attributed to Google were from this category.