Ask Siamak

Often, no — and anyone who answers this differently is selling. If you have never written down who your best customer is and why they choose you, you have a strategy problem that no amount of execution will fix, and that is worth an outside perspective. If you already know that and simply need the work done, you need an operator, not an advisor. The businesses that waste the most money are the ones that hire for execution when the underlying positioning is still unclear.

A freelancer executes a defined task — a site, a campaign, a set of posts. An agency executes across several channels with a team and overhead, and typically wants a monthly retainer. A consultant is supposed to decide what should be done and why, and the good ones stay accountable for whether it worked. The distinction that actually matters is whether the person is paid to think or paid to produce. Confusing the two is how businesses end up with excellent execution of a bad plan.

Independent consultants in the US generally run $150 to $500 an hour depending on specialization and track record; agency retainers for a small business commonly land between $2,000 and $10,000 a month. My own work is billed hourly at roughly $350 to $450, with monthly retainers for ongoing engagements. What matters more than the rate is what a customer is worth to you. A $5,000 month is expensive for a business where an order is worth $60 and cheap for one where a client is worth $40,000.

In-house wins when the work is continuous, specific to your business, and depends on knowledge that is hard to transfer — most content and most customer communication. Outside help wins when the work is periodic, specialized, or requires judgment you would only need a few times a year. The common mistake is hiring a junior generalist in-house to save money, then discovering that the hardest part of marketing is knowing what to do, not doing it.

Ask them what they would need to know before recommending anything. A good firm will ask about your margins, your customer lifetime value, your sales process, and what has already failed. A weak one will present a package of deliverables in the first meeting. Then ask for a client in your industry you can call directly — not a testimonial, a phone number. The reference call tells you more than the pitch deck ever will.

Guaranteed rankings or guaranteed placement in AI answers — nobody controls either, so the guarantee is either meaningless or being met with worthless traffic. Reporting built entirely on impressions and clicks with no line to revenue. Unwillingness to explain what they are actually doing. Long contracts with no early exit. And the quiet one that costs the most: an agency that never disagrees with you.

Not at the start. Real marketing work does take months to show results, which is the argument every agency makes for a long commitment, and it is legitimate. But you can honor that reality with a 90-day paid pilot that has defined deliverables and a clear decision point at the end. If the work is good, you will both want to continue. If a firm cannot accept that structure, they are protecting themselves against their own results.

Diagnosis before production. The first weeks should be spent finding out where you actually lose people — which is usually not where you assumed — and the output should be a written account of what is broken and in what order it will be fixed. If someone starts publishing content or launching ads in week one, they are guessing. Visible activity is easy to deliver and easy to mistake for progress.

The percentage-of-revenue rules you will find online average together businesses that have nothing in common with yours. The useful version starts with two numbers: what a customer is worth to you over the life of the relationship, and what it currently costs you to acquire one. Once you know those, the spending question mostly answers itself. I wrote the long version of this in How much should I spend on marketing.

It depends entirely on which lever you pull, and honest timelines look like this. Fixing intake — answering calls, adding a booking path, publishing the information buyers need to decide — affects revenue within weeks, because that demand already exists and is currently being lost. Paid advertising can produce results immediately and stops the moment you stop paying. Search and authority work operates on a horizon of months and compounds. Any single number offered without that distinction is a sales figure.

Work backward from money. The only questions that matter are how many qualified inquiries arrived, what fraction became customers, what each cost, and what each is worth. Everything else is diagnostic detail that helps explain those numbers. If you cannot draw a line from a marketing activity to one of them, you are not measuring, you are watching.

Those are inputs, not outcomes, and a report built entirely from them is usually hiding the fact that nobody is tracking outcomes. Impressions rising while revenue stays flat is a real and common pattern, and it means the traffic is either the wrong audience or it is arriving somewhere that fails to convert. Ask for the report that starts with inquiries and ends with closed business. If it does not exist, that is the finding.

They solve different problems and the framing is a false choice. Advertising buys you demand today and stops when the budget stops; it is the right tool for testing an offer, for seasonal pushes, and for filling a gap now. Organic search and authority are an asset you build once and keep. If the underlying offer is unproven, advertise first — it is faster and cheaper to learn from. If it is proven, not investing in the compounding asset is expensive over any real time horizon.

Referral businesses are healthier than most, and the honest answer is that you may need less than you think. But two things are worth knowing. Referred prospects almost always look you up before calling, and what they find either confirms or undermines the recommendation. And referral flow is not a growth strategy so much as a function of your existing network's size, which caps out. Marketing's job in a referral business is usually to support the referral rather than replace it.

No, but the thing people mean by SEO has changed twice. Manipulating rankings through keyword density and purchased links has been dead for over a decade. Producing genuinely useful, well-structured, findable information has never stopped working and is now doing double duty, because it is the same input that determines whether AI assistants can retrieve and cite you. The tactics aged badly. The underlying discipline did not. Full comparison: SEO vs LLMO.

Discovery of a new page takes days to weeks. Meaningful movement on a competitive commercial term takes months, and in a crowded market it can take a year. Anyone promising faster is either targeting terms nobody searches — which is easy to rank for and worthless — or doing something that will eventually cost you. Plan content strategy on that timeline or you will abandon the work right before it starts paying.

Because a search engine cannot taste your product or sit through your consultation. It ranks what it can observe: what your pages say, how clearly they say it, who references you, and what people do after they click. Being better in reality and being legible as better are two different projects, and the gap between them is fixable. Longer answer: why competitors outrank you.

You need to answer the questions your buyers ask before they buy. Whether that lives under a blog is a formatting decision. What does not work is publishing weekly on general industry topics to satisfy a content calendar — that is the version of blogging that gave the practice its reputation. Ten pages that answer real questions thoroughly will outperform a hundred that answer nothing.

Yes, but as evidence rather than currency. A reference from a source that carries weight in your field signals that you are worth referencing, and that signal now matters for AI retrieval as well as for search. Buying links, trading links, and paying for directory placement mostly do not work and occasionally cause harm. Being genuinely worth citing is slower and is the only durable version.

Nobody can guarantee a ranking, because nobody selling the guarantee controls the algorithm. The way these offers get fulfilled is by targeting a phrase so specific that you already have no competition — you rank first for something nobody searches, and the contract is satisfied. The same applies to anyone guaranteeing placement in AI answers. There is no ranking to buy and no placement to sell.

Ask it. Open ChatGPT, Claude, Gemini and Perplexity and ask each the questions your customers would ask — who to hire, where to go, what to buy — phrased the way they would phrase them. Note whether you are named, what is said about you, and whether it is accurate. That is your baseline, it costs ten minutes, and it is more useful than most paid reports.

Not today, in the sense people usually mean. There is no placement to buy inside an AI recommendation the way you buy an ad slot. What you can influence is whether models can retrieve accurate, consistent, specific information about you, and whether sources they trust reference you. Treat any vendor promising placement as a warning sign about everything else they say.

It is already reducing informational traffic — questions that used to bring people to a page now get answered in the assistant. But the visits you still receive are further along in deciding, which is why AI-referred traffic tends to convert better than general search traffic. The strategic response is not to fight for the informational clicks. It is to make sure that when the assistant answers, it names you.

It helps, and it is not the whole story. Schema makes facts explicit rather than inferred, which raises a model's confidence in stating them. But structured data describing thin or contradictory content does not help at all. The order that works is: write the facts down clearly, make them agree everywhere they appear, then mark them up.

Trace it to its source rather than trying to correct the model. Wrong information almost always originates in a stale directory listing, an outdated profile, an old press mention, or your own site contradicting itself. Fix it where it originates, make the correct version consistent everywhere, and models generally update within a few weeks as they re-crawl. There is no button that edits an AI's answer.

Not necessarily, and this is the question I most often answer with a no. Social media works when your buyers genuinely spend time there and when what you sell benefits from being seen repeatedly. For a commercial contractor whose customers arrive through referral and search, a neglected Instagram account is worse than none — it is a live signal of inattention. Being on three platforms badly is a worse position than being on one deliberately.

The one where your specific buyers already are, which is usually a shorter list than the pitch suggests. Trades, hospitality and anything visual do well on Instagram; professional services and B2B on LinkedIn; local retail and older demographics still on Facebook. The right method is not to guess from demographics but to ask ten recent customers where they spend time and how they found you.

Frequently enough to look active, which for most small businesses is once or twice a week rather than daily. The daily-posting advice comes from people whose business is content, not from people whose business is a business. A consistent weekly presence you can sustain for a year beats a daily sprint you abandon in six weeks and leaves visibly stale.

No. Purchased followers do not buy anything, they distort the audience signal the platform uses to show your posts to real people, and the inflated number is visible to anyone who looks at the ratio. Paying to promote a specific post to a targeted audience is ordinary advertising and is fine. Paying for the follower count itself is buying a decoration.

Only after you know what the account is for. Hiring someone to post without a defined purpose produces activity and no outcome, and it is one of the most common wasted line items in a small business budget. Decide first whether social exists to generate inquiries, to reassure people who already found you, or to recruit — those are three different jobs and they look nothing alike.

Usually not. Redesigns are appealing because they are visible and feel like progress, and they frequently reset hard-won search performance while solving nothing. The honest test: do you know which page loses people, and why? If not, the problem is a measurement problem, and a redesign will simply move the leak somewhere new. Fix the specific failure you can name.

Almost always one of four things: the traffic is the wrong audience, the page does not answer the question that brought them, the site does not establish enough trust to justify contact, or the path to contact you is more work than it looks. Each has a different fix, and guessing between them is why this problem persists for years. Longer treatment: traffic but no sales.

Fast enough not to lose people, which is a lower bar than the perfect-score advice implies. For a real business site running analytics, ads and review widgets, a mobile score in the 60s to 70s is respectable and the difference between 85 and 95 is rarely worth what it costs. If the page takes several seconds before anything appears, that is worth fixing immediately. Beyond that, spend the money on what the page says.

You can use it to draft, structure, and edit, and it will save real time. What it cannot do is supply the specifics that make content worth reading and worth citing — your actual prices, your actual process, the objection you hear every week, what you learned from the job that went wrong. Generic content is invisible to buyers and increasingly to models, which have no reason to cite a page that says what a thousand other pages already say.

For most businesses it remains the highest-return channel available, because it reaches people who already chose to hear from you and it does not depend on a platform's willingness to show it. The reason it underperforms is almost always the same: sending when you want something rather than when you have something. A monthly note that is genuinely useful outperforms a weekly promotion within a year.