How to Choose a Digital Agency in 2026 (5 Questions Most Won't Ask)

The right digital agency builds work that compounds — websites that don't need rebuilding in two years, brands that scale with the business, search retainers that build authority over time. The wrong agency builds work that expires before it pays back, then sells you a replacement. Knowing the difference is the difference between $50,000 well spent and $50,000 you'll spend again 18 months from now.

This guide covers five questions to ask before signing — the ones most agencies hope you won't ask.

Why agency selection matters more than ever

Digital work has more dimensions than it did five years ago. A modern marketing engagement might involve brand strategy, website build, SEO, AI tooling, conversion optimisation, multilingual readiness, and AEO — all of which need to work together as one system rather than five disconnected workstreams.

The shift means generalist agencies that ship "a website" or "an SEO retainer" without understanding how those connect to the broader stack increasingly produce work that's individually fine but collectively underperforms.

Choosing well comes down to filtering for agencies that think systemically, design for the long term, and ship work you'll still be using in five years.

Question 1: "Show me work you've built that's still being used 3+ years later"

What you're testing for: whether their work survives.

What good answers look like: specific examples — "this client launched in 2021 and is still on this site, with these updates over time, still ranking for these keywords." The agency can talk about how the work has evolved without being rebuilt.

What bad answers look like: "all our work is still being used" without specifics, or recent launches only because they only have 1–2 year track records.

Why it matters: agencies optimising for the launch sell the launch. Agencies optimising for compound returns can show you 3+ year examples of their thinking holding up. The latter is rarer and worth significantly more.

Question 2: "What's a project you turned down, and why?"

What you're testing for: whether they have judgement, or take any brief that walks in the door.

What good answers look like: specific stories — "we turned down a fintech rebrand last quarter because the brief assumed the rebrand would solve a positioning problem we didn't think a rebrand could fix." The agency can articulate what they're not good at and what doesn't fit.

What bad answers look like: generic statements about being selective, or claims that they don't turn anything down.

Why it matters: good agencies are good because they say no to bad-fit projects. If they take everything, the quality bar across their work is set by the worst client, not the best.

Question 3: "What do I own at the end of this engagement?"

What you're testing for: whether you'll be locked in to them after the work is done.

What good answers look like: specific list — "you own the code repository, hosting account, brand assets in editable formats, all CMS content, the documentation, the analytics history, the search rankings." Everything transfers cleanly if you change agencies.

What bad answers look like: "we'll send you the files" without specifics. Hosting in the agency's account. Brand assets only available as PDFs. Code on infrastructure you don't control.

Why it matters: some agencies bill less because they retain ownership of the work. The lower fee comes with vendor lock-in. Walk away from any agency that won't transfer ownership at handover.

Question 4: "How does this design for change?"

What you're testing for: whether the work is futureproof or has built-in obsolescence.

What good answers look like: specific scenarios — "the site is built modularly so you can add new sections without rebuilding," "the CMS structure scales to 10x the content without restructuring," "the AI build is model-agnostic so swapping AI providers is a config change," "the brand system flexes for additional markets."

What bad answers look like: generic claims about flexibility, or technology-led pitches ("we use the latest framework") that don't translate to durability.

Why it matters: the work that lasts is architected for change. The work that expires is architected for the launch. The difference shows up in the answer to this question.

Question 5: "What does success look like in 12 months?"

What you're testing for: whether they think in business outcomes or feature lists.

What good answers look like: measurable business outcomes — "your team is saving 8 hours per week on quoting," "your organic traffic has doubled," "your sales close rate is up 15%," "your team is publishing without calling the agency every time."

What bad answers look like: deliverable lists — "we'll have shipped the website, run an SEO retainer, produced 12 blog posts." That's input, not outcome.

Why it matters: if the agency can't translate the engagement into a business result, neither can you — and you'll struggle to justify the next engagement.

Five warning signs to walk away from

If you see any of these in the pitch, the engagement is high-risk:

• Long paid discovery before any building. 4–6 week paid discovery phases usually mean hourly billing without scope discipline.
• Pitch is mostly about technology. "We use Webflow, Next.js, GPT-5..." If technology dominates the conversation, the work is being designed around the tools, not the outcome.
• No specific examples of long-lived work. If they can't show 3+ year-old work that's still in use, the work has historically expired.
• Vague pricing. Quotes that say "starting from" without saying what the actual project will cost. Agencies that won't commit to a number won't commit to a scope.
• You'd be their biggest client. Be cautious if you'd represent more than 30% of agency revenue. Either they'll fold around you, or you'll be left in the lurch when they grow.

Five reassuring signs

By contrast, these are positive signals:

• The agency tells you when they're not the right fit and refers elsewhere
• Pricing is fixed-scope, fixed-price, with clear inclusions and exclusions
• They have a documented process for the first 4 weeks
• They can name specific platforms / approaches and explain why those, not others
• References from clients 3+ years deep into the relationship

Where to start

For most $2–10M Australian businesses, the most practical starting point is a 30-minute conversation with two or three shortlisted agencies. The conversation is more diagnostic than the pitch deck — you'll learn more from how they ask their first questions than from how they present their case studies.

For more on what futureproof digital actually looks like — and how to commission work that doesn't expire — see our guide to futureproof digital.

Frequently asked questions

How much should I expect to pay for a digital agency?

For an established Australian SMB, typical engagements: brand work $5,000–$25,000; websites $15,000–$50,000; SEO retainers $2,500–$8,000/month; AI builds $10,000–$80,000. Larger or more complex engagements scale up from there.

How long should I evaluate agencies before choosing?

2–4 weeks is reasonable. Less than 1 week and you haven't really evaluated. More than 6 weeks and you're delaying without learning more. Three serious conversations is usually enough to surface the right fit.

Should I use one agency for everything or specialists?

For most SMBs, an integrated agency that does brand + web + search (and ideally AI) produces better results than coordinating 3–4 specialists. The work compounds when one team understands the whole system. Multiple specialists end up with handoffs and gaps.

What if my budget is tight?

Tight budgets are best served by doing fewer things well rather than spreading thin. A single excellent project — a focused brand refresh, or a great website, or a serious SEO sprint — pays back better than three half-funded engagements. Pick one priority and do it properly.

Should I sign a long-term contract?

Initial engagements should be project-based with clear scope. Retainers (SEO, ongoing development) typically work on 6–12 month contracts. Avoid 24+ month commitments before you've worked together — there's no upside for you, and it usually signals the agency is locking in revenue rather than earning continued partnership.