Opinions And Ramblings By Adam Kmiec On All Things

Tag Archives: Rob Saker

Guest Post – The Impact of an Unavailable Web Site: Dennys.com Superbowl

I’m out on vacation this week.  The keys to TheKmiecs.com have been turned over to a few, select, awesome guest writers.  The following has not been edited by me and is the work and effort of the original author.  I appreciate the time and thinking that went into this post and hope you will too.  Enjoy!

It’s been a week since Denny’s ran their Superbowl ad featuring a free meal and all information seems to indicate a successful effort. The message was perfectly aligned with the economy. In a time when people are being laid off by the hundreds of thousand, offer a free breakfast to everyone in the country.

The company had seen considerable negative change in the past 20 years. They faced lawsuits accusing them of racism, and a growing public perception of inferiority, tarnishing a brand that was once widely known for good value.

Looking to put that behind them, this was an effort to reintroduce people to the value proposition of Denny’s meals. This was a big bet. Failure brings ridicule and questions on why the company spent $3 million on the effort. Success can be equally challenging, as Denny’s learned.

What Happened

Following the Superbowl and on Monday morning, Dennys.com experienced a surge in traffic with people looking for information on the offer and the location of the nearest restaurant.

According to comScore, 15% of all respondents visited an advertisers web site after seeing their web site, and 23% of those respondents visited Dennys.com. With a total audience of nearly 100 million viewers, that would place the number of visitors to Dennys.com around 3,500,000 people. Other reports vary in estimating traffic from a 434% to 1,700% spike. Both sets of figures are reasonable as the bulk of visitors likely came during peak times.

As a result of this spike, Dennys.com was largely unavailable to people after the Superbowl and Monday morning.

A Preventable Outage

It was predictable that Dennys.com would experience a significant surge following the ad. Unpredictable traffic from sources such as the DrudgeReport.com, Digg.com, and Twitter.com can lead to an unforeseen and overwhelming load on web sites, but placing a Superbowl ad requires advanced planning.

Note that Denny’s made the offer for breakfast on Tuesday morning. This may have been intended to give people time to coordinate going to Denny’s. There was clearly thought about the activation and timing of the program, but there was a either a miss in collaborating with their Digital Agency or Technology group, or a lack of experience in handling this type of event.

What Did Denny’s Miss?

It does not appear Denny’s made changes to their web site or hosting arrangement to prepare for the Superbowl traffic. A quick check of their site indicates two key misses:

  1. From a brief review of the source code and scans for origin servers, it does not appear there was a Content Delivery Network (CDN) utilized for the site. A CDN reduces the work a web server must do by offloading the amount of work it performs. A CDN such as Edgecast, Akamai, Amazon Cloudfront, or Limelight places copies of the static assets of a web site on ‘nodes’. When a user requests a web page, they are first routed to the assets on these nodes instead of the web server. In my experience, a CDN can reduce the load on a web server by over 90%, while improving response times by up to 45%.The most recent detectable change in Denny’s hosting environment occurred in October 2008 when they upgraded to a new version of IIS. We cannot tell from this information if Denny’s added capacity to their web servers.
  2. While it appears some steps were taken to minimize the number of queries required by users, the area most likely to be used by consumers was still partly dynamic (restaurant locator). Search of this type is among the most resource intensive. An alternative would have been to provide static navigation pages (state > city > locations). These could have been easily created by searching with the existing Content Management System and saving the results as static pages.

The cost to implement a CDN varies by vendor and capability, but I would estimate the total cost for both of these solutions to be less than $60,000 for the year.

The Negative Impact to Denny’s

The Internet provides us with enough information to make educated guesses on the impact of the failure. I’ve found it’s useful in the past to use a Customer Lifetime Value scenario to determine the impact of these decisions.

We first need to build a table of our assumptions:

Measure Value Method for Assumption/Calculation
Audience 100,000,000 Total potential audience of the advertisement.
Purchases per year 12 Estimated 1 trip per month.
Retention Rate per Period 70% Estimated % of customers that return the following month for meal.
Average Purchase Value $12.07 Recent breakfast ticket for two at Denny’s.
Profit Margin 20% Estimated gross profit of restaurant.
Profit per purchase $2.41 Dollarized gross profit of average ticket, calculated from Avg. Purchase with Profit Margin
Cost of Reaching a Potential Customer $0.05 Audience size/cost of advertisement.
Response Rate 2% Denny’s self-reported trial audience.
Coupon or other one-off costs $2.30 Estimated cost per meal derived from Denny’s total promotion cost.
Total Customer Acquisition Cost $2.35 Sum of per customer acquisition costs.
Technology Costs $5,000 Estimated cost per month of CDN for program of this size.

These assumptions are based on public numbers, derived from figures Denny’s has released, some educated guesses, and my personal research on Thursday at the local Denny’s. For the sake of this exercise, we’ll keep this to a one-year analysis and ignore Discount Rate and Product Inflation.

The most controversial figure in this table is the retention rate. With the exception of telecommunications, most companies do not publicize their retention rate. This guess is an average based on estimates ranging from 53 – 85% in the restaurant industry.

Based on these assumptions, we can project that Denny’s would see the following results this year:

  Q1 Q2 Q3 Q4 Total
Customers 3,066,000 1,051,638 360,712 123,724 4,602,074
Revenue $37,006,620 $12,693,271 $4,353,792 $1,493,351 $55,547,033
Cost $29,605,296 $10,154,617 $3,483,033 $1,194,680 $44,437,626
Profit $7,389,060 $2,534,448 $869,316 $298,175 $11,090,998

For a stated cost of $5 million, $11 million in gross profit is an acceptable return. It also accomplishes the goals of reintroducing the brand to consumers, and the financial impact would not be limited to 2009.

But did Denny’s leave money on the table?

Let’s next assume Denny’s implemented the recommended technology steps at an additional cost, and as a result saw an increase of 5% in the response rate to the trial (to a total of 2.1%). What impact would that have on the overall program?

  Q1 Q2 Q3 Q4 Total
Customers 3,219,300 1,104,220 378,747 378,747 4,832,178
Revenue $38,856,951 $13,327,934 $4,571,481 $1,568,018 $58,324,385
Cost $31,113,438 $10,681,764 $3,673,700 $3,673,700 $46,659,508
Profit $7,743,513 $2,646,170 $897,781 $298,084 $11,664,877

The most immediate impact is Denny’s would have served 153,300 more customers on February 3. If we follow the same assumptions through the year and add the costs for the technology recommendations Denny’s would have realized an additional $573,879 in profit for the year. That is a 956% return on the technology investment.

Note: These figures also raise an interesting question into what type of retention programs they implemented. With retention driving significant dollars, was the bigger mistake to not leverage CRM?

What does this tell us?

This wasn’t a Grand Slam for Denny’s, but it was a solid double. Denny’s failure here isn’t critical from a financial perspective. The program will likely achieve profitability for the restaurant, and bring new customers into the stores. The idea with any trial promotion is to introduce people to your brand with the hopes a percentage of the trial audience will return and by those definitions, the company was modestly successful.

While the company will benefit from its efforts, I think it stumbled seriously with the online execution. There was no excuse for Dennys.com to be unavailable to users. For a company that has a reputation issue, not being able to serve people online is a failure. They had the advanced notice. Technology is readily available to serve capacity, and they were clearly thinking about the timing.

Located in Chicago, Rob Saker is a Marketing Technology & Analytics professional in the Consumer Goods industry. He can be reached at his blog at www.robsaker.com.