When search ads are worth it
Work backwards from what a customer is worth, not forwards from what a click costs. If that sum does not work, nothing else in the account will save it.
Search advertising has one property nothing else has: you can put an offer in front of somebody at the exact moment they are looking for it. That is worth paying for. It is also why the prices are what they are, and why plenty of businesses lose money doing it competently.
Before anything about accounts, keywords or bidding, there is one question. Can you afford a customer at the price this channel charges? Everything else is detail on top of that answer.
The only sum that matters at the start
Work backwards from what a customer is worth, not forwards from what a click costs. Three numbers, and you can estimate all of them.
- What a customer is worth to you, over the whole relationship, in profit rather than revenue.
- What share of the people who land on your page become customers. If you have no data, guess low and plan to find out.
- What you can therefore pay for a click: the first number multiplied by the second, then cut to leave yourself a margin.
As an illustration, not a benchmark: if a customer is worth 300 in profit and one visitor in fifty buys, each visitor is worth 6 on average. Paying 2 a click leaves room. Paying 5 does not, once you account for the months where the rate is worse than you assumed.
The same sum with a longer relationship
Services usually have a second and third purchase in them, and leaving that out is the most common reason the channel looks unaffordable when it is not. Illustrative figures again.
Say a first project is 900 and your costs on it are 300, so 600 in profit. Say that roughly one client in three comes back for a second project in the following year. The expected profit per new client is therefore closer to 600 plus a third of 600, around 800, rather than 600.
If one enquiry in four becomes a client, an enquiry is worth around 200. If one visitor in twenty enquires, a visitor is worth about 10. At a click price of 3 you have room; at 8 you do not, unless the conversion rates are better than you assumed, which at the start they will not be.
Do this sum on paper before you open an account. It takes ten minutes and it is the difference between advertising and donating.
What to assume when you have no data at all
You need a number to start and any number you invent will be wrong. Two principles make the wrongness survivable.
- Assume worse than you hope, then treat the first month as buying the real figure rather than as earning anything. If the sum only works at your optimistic rate, you do not have a plan, you have a wish.
- Measure the step closest to money that you can see. For a service that is enquiries, not sales, because sales depend on your closing as well as the advert. Keep them separate or you will blame the channel for something you did.
The budget has to be big enough to teach you something
This is the point most often missed, and the arithmetic is unforgiving. Illustrative: at 3 a click, 300 a month buys a hundred clicks. If one visitor in twenty enquires, that is five enquiries a month across the whole account.
Five enquiries cannot tell you which of your six keywords works. It cannot tell you whether a change helped. You will stare at it, see noise, and turn it off in week three having learned nothing except that you spent 300.
The way out is not more money, it is less surface. One campaign, one tight group of keywords, one page. The same 300 spent on four keywords instead of forty produces enough events on each to mean something. A budget too small to produce data is not a cautious start, it is spending money without buying information.
When it makes sense
- People search for what you sell using words that describe buying it, not just learning about it.
- A customer is worth enough to absorb a few clicks that go nowhere, because most of them will.
- You can take the money without a long sales process. Ads for things that need six meetings can work, but the feedback loop is slow enough that you will be guessing for months.
- You can tell which clicks turned into money. Without that, you are not advertising, you are donating.
When it does not
- Nobody is searching yet. If the problem has no name people type, demand has to be created elsewhere first, and search is the wrong tool.
- The margin is thin and the price is set by competitors. You will be outbid by whoever can afford the thinnest margin.
- You have no page worth sending traffic to. Paid traffic hitting a weak page just buys you a faster answer to a question you could have asked for free.
- The budget is small enough that you will get a handful of clicks a week, across more keywords than that handful can inform.
On bidding for your own name
A question that comes up early. Someone searching your business name already knows you, so paying for that click can feel like paying for traffic you would have had anyway, and often it is.
It is worth doing in two cases: when a competitor is bidding on your name and taking the top of the page, and when the results for your name show something you would rather they did not. Otherwise it is cheap traffic that flatters your account's average without adding customers, and it is worth knowing that is what it is doing.
Work to do
- Write down what a customer is worth to you in profit, not revenue, including any repeat work.
- Write down what share of visitors become customers. If you have never measured it, halve your first guess and write that instead.
- Multiply them, take off a margin, and compare the result to what a click actually costs for your main keyword.
- Divide your monthly budget by the click price. If the result is fewer than about a hundred clicks, cut the number of keywords until the clicks per keyword could teach you something.