Brian David Hall

Agency Optimization

Don't Automate Yet

22 November 2018

When processes are broken or inefficient and client deliverables aren’t getting … delivered … it’s tempting to look to automation as a solution. “This shouldn’t be so hard” you’ll likely think. “We can find, or build, a tool to do this for us; then it’ll always get done.”

This dreamkiller post is meant to dissuade you from pursuing that approach. We’ll look at the very reasonable motivation to automate, examine the costs of automating, and prescribe a different solution to your immediate problem.

Why it’s tempting to automate

It’s likely about feelings. Missing deadlines or failing to deliver is frustrating and embarrassing. Maybe it even makes you angry. After all, it’s your agency and ultimately your responsibility to see that all obligations are met.

Automation offers the promise of never having to feel this way again. Set it up once, push a button, and rest easy in the knowledge that everything’s handled, forever and ever.

As an example, suppose a client expects to receive a marketing report every month. For some reason, it didn’t go out last month. Maybe it hasn’t gone out for a few months, and now they’re upset. And now it’s your problem. Who wouldn’t want to solve this problem forever with the help of a trusty 🤖? Can’t you just set up a report to go out automatically on the first of the month, add the client’s email address, and be done with it?

Disappointing a client, and having to drop everything to fix it, is costly. It hurts your chance of renewal, your reputation, your productivity, and your mood. The urge automate a solution to solve it now and forever is a healthy and wise response, one that recognizes the larger problem behind the current situation. But … don’t do it.

The costs of automation

First and most obviously, an automated solution costs time up front to set up and configure. You already knew this, though, and are likely more than willing to invest that time.

The real risk here, the monster under the bed that should terrify you, is the risk of automating an incomplete or inappropriate solution. And if you’re seeking automation to solve an emergent problem, this risk is significant.

Going back to the client reporting issue, you might set up the report to go out automatically. You can be confident that they’ll get one at the beginning of next month. But if they haven’t received anything in two months, how confident are you that you’re sending them what they need? Would they also appreciate a (tragically not automatable) executive summary? Will this report raise more questions than it answers?

The underlying problem is that you’re automating a process that either doesn’t exist yet, or isn’t quite working yet. When this happens, not only does the original probelm persist, new problems arise. Your client was hoping for a tennis pro to give some pointers on her serve, and instead you’ve pointed a tennis ball machine at her, pressed ‘Start’, and walked away.

The dangers of premature automation

Solve it with people and processes

Every great automated solution starts with a checklist. So make one. Don’t just make one up, though - step through a manual solution to your problem and make one as you go.

The harsh reality is that you’ll have to solve your problem by hand at least a couple times before you can hope to automate it. The upside is that you’ll never be back in this situation again.

The lure of automation causes us to transpose what’s needed to what’s automatable, and accept that as a solution. “The client needs a report” becomes “Reports are automatoble” and then “We’ll send the client an automated report.” What gets lost in that process are the specifics of what the client needs.

Solving the problem by hand forces you to keep this goal, your true goal, in mind throughout the process. As you build your checklist, you’ll probably notice a few decidedly non-robotic steps. “Add summary of campaign strategy for next month” or “Call out performance metrics for the Philadelphia branch.” Don’t worry that you can’t push-button these steps - you’re delighting the client and saving yourself future grief. You couldn’t have automated this part, so your robo-report would’ve missed it. And you won’t have to do it forever.

Once you’ve shipped the report, set a calendar event for the first of next month to generate the report again, and invite a virtual assistant, or an employee to whom you can hand off this task.

When that meeting comes, step through your checklist and create the report. You’ll probably notice a bit of missing detail or some changes to the process as you go through it. Congratulations! Celebrate your wisdom for not automating the wrong process. Once your new reporting expert is clear on the process, set a calendar event for the second of next month to check in on him.

At the next meeting, confirm that the report went out. If it did, great - you’ve solved the problem you set out to solve with automation. Now, ask: how long did it take? Did you get stuck anywhere? Was any part of the task ambiguous or confusing?

The answers to these questions will point to opportunities to automate parts of the process. But they’re just as likely to point to the fact that there’s no need to automate it. If a trusted resource can complete the task in 30 minutes a month, you’re done. If there’s a true economic case to increase efficiency here, you can hand the relevant checklist steps to a developer and count on having it built right. (Tell a dev to “go automate reporting” and there’s no telling what you’ll get. Tell her to “produce an automated report on total spend, impressions, and CPC for Philadelphia, to be uploaded to Google Drive on the first of each month” and you’re in business.)

So if you’re dealing with fallout from a broken process, don’t look to a 🤖. Look to a checklist, a calendar, and a human for help.

There’s no getting out of the need to roll up your sleeves and solve complicated client problems; that’s what they count on you for. But you only have to roll up those sleeves once or twice. And the processes you build by doing this add directly to your agency’s value - they increase output, they remove you from the day-to-day, and they help teach your team how to solve their own problems. Whether your long term goal is to sell the agency, or just take a vacation, it’s checklists ✅ not robots 🤖 who will make it possible 😁.


Photo by Rock’n Roll Monkey on Unsplash