Starting from Scratch: A Paid Search Primer
We’ve talked a lot about how to get the most out of your Paid Search program, as well as how to get the most bang for your buck from your SEM. But what if you’ve just decided to venture into the paid search arena? How do you get started marketing your products via paid search?
As George pointed out in his post last week (PPC Outsource or In-House? Rules of Thumb), using publicly available tools — such as a combination of Google Analytics and Adwords reporting — make a lot of sense if you’re just getting started. These resources, coupled with a little knowledge of manipulating data in an Excel spreadsheet, can go a long way in managing a fledgling program. Here are a few high-level strategies to keep in mind when launching a PPC program from scratch:
- Set specific goals. Even before you start running paid search ads, ask yourself, “What do I want the program to do?” and “How will I measure these goals?” Your goal could be email signups, registrations or (most likely) converting clicks directly to orders and sales.
- Set specific economic targets. Once you’ve established overall goals for the program, determine what you’re willing to pay to drive toward these goals. Select specific cost targets that will help you to meet the overall goal for the program. Typically, it’s easiest to focus on a cost-per-lead, cost-per-order or a cost-per-sales-dollar. These financial targets should be set with your current acquisition costs in mind. We’d recommend setting your targets at or below your current acquisition costs. For example, if you typically spend $20 in marketing and advertising to obtain an order with average value of $100, you might consider setting your PPC goals to meet or exceed that same ratio: an average cost-per-order of $20 or an ad-cost-per-sales-dollar of 20% ($20/$100).
- Start with Google. Google commands the lion’s share of the search traffic. Focusing your efforts here will allow you to focus more on strategy and gather more data in a shorter period of time. It will also help in controlling budgetary concerns. General lessons learned can them be applied to other engines. (Though they will behave differently at the keyword level.)
- Test the waters with your brand. If you’ve had a catalog or web presence for a while, and you don’t have a high-ranking natural result for your brand name, perhaps consider starting with a few brand terms. It’s important to note that folks coming in on your brand aren’t necessarily new customers—they were searching for you anyway. Likely, results on brand terms will far exceed any economic targets that you have set out for your program. Brand terms just make it easier for customers to find you, shop, and place an order online. You likely won’t see the overall sales needle move, however you may see some variance in your sales channels—from phone to web, perhaps.
Because clicks on brand ads are typically representative of customers who were already searching for you by name, it’s important that you consider the success of these types of terms separate from terms that don’t include your brand.
- Be specific when adding non-brand terms. Resist the urge to create general keywords with the intention to capture a lot of traffic. “Testing” a few general search terms on the first page rarely works. General terms typically cost more and you’re likely to lose your shirt in the process.
When you’re creating your terms, it’s important to remember that search is a ‘pull’, not a ‘push’ medium. It’s unlikely that you’re going to create demand for a product that no one knows about. On the flip side, using specific terms that describe specific products will allow people to find you. Start with your most popular category or group of products and develop 2-3+ specific keywords for each product. (Eg: keywords that contain “Model + SKU”, “Brand + Model” and “Style + Brand + Model”) Keep in mind how people search. Mining your site-search logs can give some indication to how folks look for your products. Be sure to only create terms for products you carry.
By starting with a limited selection of specific terms, you’re effectively starting at the proverbial long tail of search, with the idea of moving toward more general terms, if these are successful. This approach is most easily implemented with hard-line products where the consumers are very SKU and model savvy (think consumer electronics). Although it may be slightly more challenging to create the terms, the same concept can be applied to clothing, food products, etc.
On an individual level, these specific, long tail terms may be very low traffic, but they’re also going to be low cost and high converting. This approach also helps you to better understand how the public searches for your product. After a month or so, you may feel comfortable adding more general terms for the same category (think a level up—“Model + Product” , “Model”, and “Product”). Be sure to start these general terms out on phrase match or exact match to ensure that these ads aren’t served for any overly-broad searches.
If you don’t have it, don’t advertise it. Be sure you’re only adding terms for specific products you have and only add general terms if you have a good selection of products matching that description. (eg: Don’t run the term ‘GE Toaster’ if you only sell one hot-pink model. Stick with specific terms like ‘hot-pink GE toaster’ and ‘GE Toaster 169T81 pink’. These terms may not drive a lot of traffic, but you’ll be there when someone is searching for the perfect toaster for their hot-pink kitchen.)
Be cautious with budget caps. It’s smart to have a budget for your paid search program. It’s also smart to manage that budget through term selection and bidding. Use some budget caps in the first week or two, till you get an idea of the amount of ad spend you’ll see each week. Using budget caps as a long term strategy will increase the likelihood that you overspend on non-producing terms and shut off ads before more specific high-converting terms have a chance.
If you’re budget-constrained and hitting the daily budget caps, reduce bids so that you’ll capture more clicks for the same total dollar spend. Keep dropping the bids until the budget caps are no longer a factor. This will get you the most sales for your limited budget.
Have patience. It may take a while for your specific, long tail terms to get a click. Although you should watch program carefully and react quickly to what you see, resist the urge to add too many general terms too quickly. It could take 60-90 days to get an assessment of what’s happening, especially if you started with a small, specific term list.
Conversely, don’t turn off under-producing terms too quickly. It’s not a hard and fast rule, but you can use the average order value for the category of a keyword (or group of keywords) as a proxy for how much to spend before considering that keyword to be a dog.
Don’t stress about click-to-order windows. When you’re getting started, consider crediting any order that comes in through PPC to the program—don’t worry about limiting your tracking window in the first 60-90 days. Once you have a full selection of keywords running, you may want to consider limiting the time from ad click-to-order. Typically 30-60 days is enough time for highly considered purchases (eg: high dollar items, electronics) while 15-30 days may be more appropriate for categories like apparel and drugstore items.
Re-evaluate budgets and economic targets. If the non-brand keywords in the program are successful—exceeding your economic targets—consider increasing the paid search budget. The concept here is ‘If every time you hand me ten dollars, I hand you back eleven, why would you ever stop handing me dollars?’ Remember, the success of the paid search program should be based on your non-brand keywords. Be sure you’re regularly auditing the program for economic efficiency.
Once you’ve assessed this intro program, continue to build out specific keywords for your other categories. Adding terms is a constant process of testing to see what works and what doesn’t. If your term list becomes unmanageable with your current tools and you’re still driving significant traffic at a rate that exceeds your economic expectations, it might make sense to consider an outsource solution.