You have to work with the clients you can get, not the ones you hope for. Sometimes - especially in the beginning - you might not be able to attract big-budget clients.
I accidentally ended up owning a Google Ads agency where 99% of my clients were spending < $300 a month on ads. (How it happened is a story for another time.)
I’ve been lucky enough to work with bigger clients since then, but, those first low-budget clients got me there. I wouldn't have been able to stay in business long enough to move up the ladder without them.
Update: You might find this article on growing your own bigger-budget clients useful
Here's a list of things I wish someone had told me in my first year of working with low-budget clients.
Say no to the wrong kind of clients. Horrible people. People who don't have the skills to use the internet. People who need a follow up before they reply to your email.
Play Google Ads on the easy level. Low budget clients can’t afford enough clicks, or enough of your time, to compete in tough markets. Forget car rental, financial services or general legal.
Don't be anyone's last hope. It's almost impossible to help a client who has to make sales NOW or go under. There isn't time to ramp up a campaign, get some leads and teach the client to sell to strangers on the internet. Far better to deal with stable businesses that want to grow a little.
Automate low-value work. Google has built in automation - bidding, rules, scripts and an API. 3rd parties offer even more ways to get out of routine, low-value work.
Specialise to get the economies of scale. I do lead generation on Google Ads. No ecommerce, no Facebook, no Instagram. You might choose to specialise in a niche - dentists, balloon tour operators etc - instead. Doing the same kind of campaigns over and over makes you faster and better.
Batch your work. Doing the same kind of work for all campaigns is more efficient than working one campaign at a time. For example check the search terms report for negative keywords for all campaigns. Then move onto checking split test ads for all campaigns. It'll feel easier than doing different tasks one campaign at a time.
Check for problems automatically. You shouldn't be checking 300 landing pages every day by opening each one. You shouldn't have to log into 50 accounts every morning to check for disapproved ads. This can be automated easily.
Stay out of the account unless you’re doing scheduled work or fixing a problem. Popping in to check how things are going steals time.
Use your own landing pages not the client’s site. It's faster and easier than chasing the client for access credentials, learning how their home-made CMS works or wrestling with their web developer to get conversion tracking working or a hostile contact form fixed. Also, most client websites aren't designed to convert.
Don’t allow exceptions to your business processes. I have regretted every exception I’ve ever made to our standard business processes. At some stage you will forget that you’ve made that exception. Forgetting will cause a problem that will cost a disproportionate amount of time, and possibly money, to fix.
Use checklists for things like taking on a new client or building a new campaign.
Systemise. Having set processes and systems allow you and your staff to do your best work every day - not just on your best day. I'm influenced by the book The eMyth Revisited.
Write template emails. Save your replies to client emails. Come back later and polish them into a template that you can copy and paste the next time the question comes up.
Set expectations up front. What's included in your service. More importantly, what's not included.
Teach your clients to sell. I send this article to all new clients.
Watch for the few clients who take up most of your support resources. The 80/20 rule applies to client support requests. 80% of your clients won’t ask for much. A small proportion will bleed you dry.
Block out time for incoming client calls. I take incoming client calls on Tuesday and Thursday afternoons. The rest of the week is mine.
Allow clients to book their own calls. Email ping-pong is a waste of time. I use Calendly for this.
Pre-empt common questions. How often have you been asked "Why can't I see my ads?" I pre-empt this and other common questions with a series of emails for new clients.
Don’t build everything for small campaigns. A small budget can't support every feature and opportunity Google offers.
Focus on bottom of funnel keywords. These will lead to quicker sales.
Don’t split test all your adverts. Split test in your busiest ad groups and roll winning elements out everywhere.
3 bullet-proof optimisations will get you most of the way there. Police search terms reports for negative keywords. Promote high-converting search terms to an alpha campaign with it's own budget. Block low-converting search terms. (Update: This was correct at the time of writing but, given how frequently things change, it might not be the best way to optimise a campaign today.)
Reports invite questions. Reporting numbers without context for those numbers is meaningless.
Don’t report on your activity. You wouldn't make changes without enough data. You wouldn't stop a split test before there was a result. Sometimes doing nothing is best for the campaign. But, sending a report full of nothing makes it look like you've been lazy. Better not to send these kinds of reports at all.
Don’t report on intermediate metrics. Metrics like CTR and QS are there to help you, not the client.
Report on metrics closest to the client's business. I send one report, automatically, every Friday. It contains a list of leads generated this week. Nothing else.
Complex reports get archived, unread, after the second month.
Charge up front for setup. Amortising the cost of a new campaign build in your monthly invoice makes your fee look huge compared with the ad budget.
Don’t price as a percentage of ad spend. Percentage based pricing works for medium budgets. It breaks down at low and high budgets. I prefer a fixed monthly fee.
Don’t offer credit. You need a credit control department if you invoice in arrears. You have to be able to assess client's creditworthiness, issue invoices, chase late payers and take action when people don't pay. Better to avoid this and get your money up front.
Have your clients pay Google. See above.
Automate monthly billing. Stripe bills my clients every month. If the charge is successful, Stripe tells our system. Our system generates an invoice, marks it paid, turns it into a PDF and emails it to the client. If there is a problem with the payment, Stripe tells our system and the system tells the client. There is no human intervention.