What Every Multichannel Merchant Should Know About Paid Search Performance Audits

As a cataloger/multichannel marketer, you’ve long understood the importance of double-checking all aspects of your marketing programs to make sure everything is in order. No doubt, you visit your printer when you’re on press to monitor print quality. You do bindery checks to inspect book assembly. You likely use mail decoys to confirm delivery. You also probably “mystery shop” your own company to monitor your call center and shipping teams. You surely double-check your printing and postal invoices for accuracy.

The same attention to detail applies to online marketing. As paid search marketing grows in importance and consumes a larger share of catalogers’ acquisition budgets, it’s prudent to scrutinize these programs regularly. You want to make sure your plans are being executed correctly, and that your precious marketing dollars are being used intelligently.

Here are my recommendations for auditing paid search marketing campaigns. I suggest performing these checks at least quarterly. It might take some calls and e-mails to assemble the data you’ll need, but once those data are in hand, this audit should take less than a full workday. For the time period, perform your audit on last month’s complete data.

Scrutinize Sales Data

1. Make sure your sales data are accurate. The goal of this step of the audit is to confirm that the sales reported by your search marketing reports are accurate and that you’re not overcounting, undercounting or double-counting orders. If you aren’t tracking sales from your paid search programs, you must — plain and simple. Buying clicks without watching corresponding conversions is like flying an airplane blindfolded: It’s not a matter of if you will crash, just a matter of when.

If you have several systems tracking sales, determine which of them you’ll consider as authoritative. The various systems will certainly differ — often significantly. For instance, say your authoritative paid search sales report indicates $1 million in paid search revenue for last month. Pull the corresponding order numbers. Make sure those orders actually total $1 million. Then check each order against your back-end accounting system and ask the following questions:

* Did these orders actually ship?

* Were these orders attributed to paid search correctly, or should some be assigned to another marketing source?

If you find anomalies, dig in deep and “pull the thread” until you understand the discrepancies.

2. Make sure your cost data are accurate. The goal of this step is to confirm that the costs in your search reports match reality. Follow these steps:

* Take the total costs you paid last month to the search engines (by engine) from your search marketing reports.

* Confirm that the amounts match the actual
monthly invoices from the engines if you or your agency are invoiced for your clicks; or match the actual credit card charges if you pay by credit card.

* Finally, check both sets against the engine’s online accounting reports.

* You should have log-ins to all your paid search accounts. Make sure that all three sets of costs — as reported in your search reports, in the engine’s online accounting applications and those costs actually paid by your accounting department — all match up.

* Scrutinize term-level performance for your most important terms to see if your term-level economics make sense.

Build Spreadsheets

3. For each engine, pull a list of your top 100 phrases, ranked by cost and (separately) by sales. Create a 200-row spreadsheet for each engine with these columns:

* phrase,

* impressions,

* clicks,

* ad cost,

* resulting orders,

* resulting sales and

* average position.

If you end up with fewer than 200 total rows, with some terms appearing on both lists, don’t worry. You should be able to access clicks, cost, sales, and position data trivially. If you can’t, you need better tracking or better analytics or a new search agency. Add these computed columns:

* CPC — “cost per click” — ad cost divided by clicks;

* CPM — “cost per 1,000 impressions” — ad cost divided by impressions multiplied by 1,000;

* CTR — “clickthrough rate” — clicks divided by impressions;

* CONV — “conversion rate” — orders divided by clicks;

* SPC — “sales per click” — sales divided by clicks; and

* A/S — “advertising-to-sales-ratio” — advertising divided by sales.

Now sort this sheet by descending cost. Are your most costly terms appropriate for your business? Scrutinize generic terms closely. Generic words like “gift,” “coat” or “travel” make no sense for catalogers concerned with profits.

Next, are your most costly terms carrying their weight? Check their ad to sales (A/S) figures to ensure they’re selling enough to justify their costs.

Now, sort this sheet by ascending A/S, bringing your highly efficient terms to the top of the page. These should have low average positions. Low position means higher on the search results page, indicating more prominence and more traffic.

If you find low A/S terms with high position, you’ve been underbidding some winners and missing out on potential sales. If you find high A/S terms in low positions, you’ve been overbidding some loser terms.

Perform all of these spreadsheet analyses separately for each engine, as the engines differ in their performance. Separate “brand” from “nonbrand” results. The goal of this step is to understand your search economics beyond your brand halo.

Gauge Search Word Origins

4. If you’re an established cataloger with a well-known brand (think Lands’ End, L.L. Bean, Harry & David, J. Jill, etc.), a large portion of your pay-per-click sales will come in on your brand name, your brand name plus additional words (“Bean boot”), misspellings (“Hary & David”) and URLs.

These are great terms enjoying high sales at a low cost, but the sales likely aren’t incremental. When searchers type your brand into a search engine and then follow up by clicking on a paid search ad tied to your brand, they’re using the search engine like a White Pages lookup (“Get me to Lands’ End!”) vs. a Yellow Pages lookup (“Where should I buy a blue men’s oxford shirt?”). This traffic reflects your brand equity built up from years of catalog mailings — from positive word-of-mouth, magazine ads, returning customers and so on.

It’s important to segregate all these brand terms from your nonbrand results. Your portfolio of non-brand terms represents your acquisition efforts, bringing in customers who weren’t looking specifically for you.

The efficiency will be worse in your nonbrand portfolio. That’s normal and it’s OK to use the high-sales/low-cost branded portfolio to subsidize the nonbrand portfolio by a reasonable degree. What you don’t want to see is truly horrible nonbrand terms hiding behind the positive halo of your branded terms.

Hit the engines and check some searches. The goal here is to view your campaigns though a potential customer’s eyes.

5. Put aside your invoices and spreadsheets, and do some searching on Google and Yahoo! Do your ads show up for searches on your brand or your seasonal specials or your detailed SKUs, your manufacturer’s brands and item names? Note any “search holes.” Either your ad isn’t appearing for economic reasons (check those spreadsheets again) or you’ve found a gap in your term list.

For each ad, is the copy detailed, focused and compelling? Do any of the ads under- or overpromise? Following the click, do you reach a relevant deep page? Are the destination URLs as tightly matched to the search phrase as possible? For example, “men’s hiking boot” should reach a page of men’s hiking boots, not a page offering boots for men, women and kids.

The Big Picture
By the end of this day, you’ll have a good sense of the health of your paid search programs. Search is complex, so you’ll likely find some minor hiccups. That’s fine; just ask your team or agency to fix those.

Just hope you never find programs running on shaky cost or sales data, programs burning large amounts of ad costs without corresponding sales, or programs with sloppy copy or landing pages.

Bringing previously hidden problems into the bright sunlight is never fun, but it does provide the starting point to get your programs into tip-top shape. May all your audits be unsurprising, and may all your clicks sell.

Alan Rimm-Kaufman is CEO of the Rimm-Kaufman Group, an online agency helping catalogers manage paid search and improve site conversion. You can reach him at (434) 970-1010 or at www.rkgblog.com.

For related articles, visit the E-commerce Community.